pragent feedback roll-up #8

New Issue

Open

opened 2026-08-20 19:58:34 +00:00 by pragent-bot · 18 comments

pragent-bot bot commented 2026-08-20 19:58:34 +00:00

Copy Link

Copy Source

pragent feedback roll-up — auto-created.

pragent feedback roll-up — auto-created.

pragent-bot bot commented 2026-08-20 20:08:33 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-08-20 20:08 UTC

• findings analyzed: 0
• repos with feedback: 0 ()
• case-review queue: 0 disagreement(s)

Restraint

✅ Bot stayed quiet on 100% of reviewed PRs (0 / 0). Restraint OK.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 0 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

none yet

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-08-20 20:08 UTC - **findings analyzed**: 0 - **repos with feedback**: 0 () - **case-review queue**: 0 disagreement(s) ## Restraint ✅ Bot stayed quiet on **100%** of reviewed PRs (0 / 0). Restraint OK. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 0 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. _none yet_ _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-08-21 03:07:12 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-08-21 03:07 UTC

• findings analyzed: 29
• repos with feedback: 2 (gitea_admin/pragent, netcracker/interview)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 83% of reviewed PRs (5 / 6). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 0 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

none yet

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-08-21 03:07 UTC - **findings analyzed**: 29 - **repos with feedback**: 2 (gitea_admin/pragent, netcracker/interview) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **83%** of reviewed PRs (5 / 6). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 0 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. _none yet_ _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-08-22 03:07:08 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-08-22 03:07 UTC

• findings analyzed: 29
• repos with feedback: 2 (gitea_admin/pragent, netcracker/interview)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 83% of reviewed PRs (5 / 6). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 0 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

none yet

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-08-22 03:07 UTC - **findings analyzed**: 29 - **repos with feedback**: 2 (gitea_admin/pragent, netcracker/interview) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **83%** of reviewed PRs (5 / 6). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 0 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. _none yet_ _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-08-23 03:07:13 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-08-23 03:07 UTC

• findings analyzed: 29
• repos with feedback: 2 (gitea_admin/pragent, netcracker/interview)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 83% of reviewed PRs (5 / 6). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 0 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

none yet

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-08-23 03:07 UTC - **findings analyzed**: 29 - **repos with feedback**: 2 (gitea_admin/pragent, netcracker/interview) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **83%** of reviewed PRs (5 / 6). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 0 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. _none yet_ _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-08-24 03:07:59 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-08-24 03:07 UTC

• findings analyzed: 32
• repos with feedback: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 47% of reviewed PRs (7 / 15). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 0 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

none yet

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-08-24 03:07 UTC - **findings analyzed**: 32 - **repos with feedback**: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **47%** of reviewed PRs (7 / 15). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 0 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. _none yet_ _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-08-25 03:07:19 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-08-25 03:07 UTC

• findings analyzed: 32
• repos with feedback: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 47% of reviewed PRs (7 / 15). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 0 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

none yet

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-08-25 03:07 UTC - **findings analyzed**: 32 - **repos with feedback**: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **47%** of reviewed PRs (7 / 15). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 0 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. _none yet_ _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-08-26 03:07:09 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-08-26 03:07 UTC

• findings analyzed: 32
• repos with feedback: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 47% of reviewed PRs (7 / 15). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 0 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

none yet

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-08-26 03:07 UTC - **findings analyzed**: 32 - **repos with feedback**: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **47%** of reviewed PRs (7 / 15). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 0 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. _none yet_ _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-08-27 03:07:11 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-08-27 03:07 UTC

• findings analyzed: 32
• repos with feedback: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 41% of reviewed PRs (7 / 17). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 0 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

none yet

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-08-27 03:07 UTC - **findings analyzed**: 32 - **repos with feedback**: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **41%** of reviewed PRs (7 / 17). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 0 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. _none yet_ _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-08-28 03:07:06 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-08-28 03:07 UTC

• findings analyzed: 32
• repos with feedback: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 41% of reviewed PRs (7 / 17). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 0 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

none yet

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-08-28 03:07 UTC - **findings analyzed**: 32 - **repos with feedback**: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **41%** of reviewed PRs (7 / 17). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 0 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. _none yet_ _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-08-29 03:07:08 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-08-29 03:07 UTC

• findings analyzed: 32
• repos with feedback: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 41% of reviewed PRs (7 / 17). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 0 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

none yet

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-08-29 03:07 UTC - **findings analyzed**: 32 - **repos with feedback**: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **41%** of reviewed PRs (7 / 17). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 0 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. _none yet_ _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-08-30 03:07:07 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-08-30 03:07 UTC

• findings analyzed: 37
• repos with feedback: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 46% of reviewed PRs (11 / 24). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 0 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

none yet

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-08-30 03:07 UTC - **findings analyzed**: 37 - **repos with feedback**: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **46%** of reviewed PRs (11 / 24). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 0 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. _none yet_ _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-08-31 03:07:13 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-08-31 03:07 UTC

• findings analyzed: 37
• repos with feedback: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 46% of reviewed PRs (11 / 24). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 0 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

none yet

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-08-31 03:07 UTC - **findings analyzed**: 37 - **repos with feedback**: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **46%** of reviewed PRs (11 / 24). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 0 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. _none yet_ _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-09-01 03:08:53 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-09-01 03:08 UTC

• findings analyzed: 48
• repos with feedback: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 47% of reviewed PRs (18 / 38). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 0 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

none yet

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-09-01 03:08 UTC - **findings analyzed**: 48 - **repos with feedback**: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **47%** of reviewed PRs (18 / 38). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 0 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. _none yet_ _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-09-02 03:07:23 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-09-02 03:07 UTC

• findings analyzed: 51
• repos with feedback: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 48% of reviewed PRs (20 / 42). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 1 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

AC	repo	path:line	sev	problem	votes	resolved	seen
1	gitea_admin/pragent	pilot/review/opencode_lens_config.py:58	INFO	🟡 [MEDIUM] _coerce_str is defined twice in this module — lines 29-30 and lines 58-59. The second de…	👍0 👎0	✅1	1

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-09-02 03:07 UTC - **findings analyzed**: 51 - **repos with feedback**: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **48%** of reviewed PRs (20 / 42). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 1 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. | AC | repo | path:line | sev | problem | votes | resolved | seen | |---|---|---|---|---|---|---|---| | 1 | `gitea_admin/pragent` | `pilot/review/opencode_lens_config.py:58` | INFO | 🟡 [MEDIUM] _coerce_str is defined twice in this module — lines 29-30 and lines 58-59. The second de… | 👍0 👎0 | ✅1 | 1 | _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-09-03 03:07:16 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-09-03 03:07 UTC

• findings analyzed: 51
• repos with feedback: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 48% of reviewed PRs (20 / 42). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 1 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

AC	repo	path:line	sev	problem	votes	resolved	seen
1	gitea_admin/pragent	pilot/review/opencode_lens_config.py:58	INFO	🟡 [MEDIUM] _coerce_str is defined twice in this module — lines 29-30 and lines 58-59. The second de…	👍0 👎0	✅1	1

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-09-03 03:07 UTC - **findings analyzed**: 51 - **repos with feedback**: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **48%** of reviewed PRs (20 / 42). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 1 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. | AC | repo | path:line | sev | problem | votes | resolved | seen | |---|---|---|---|---|---|---|---| | 1 | `gitea_admin/pragent` | `pilot/review/opencode_lens_config.py:58` | INFO | 🟡 [MEDIUM] _coerce_str is defined twice in this module — lines 29-30 and lines 58-59. The second de… | 👍0 👎0 | ✅1 | 1 | _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-09-04 03:07:13 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-09-04 03:07 UTC

• findings analyzed: 51
• repos with feedback: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 48% of reviewed PRs (20 / 42). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 1 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

AC	repo	path:line	sev	problem	votes	resolved	seen
1	gitea_admin/pragent	pilot/review/opencode_lens_config.py:58	INFO	🟡 [MEDIUM] _coerce_str is defined twice in this module — lines 29-30 and lines 58-59. The second de…	👍0 👎0	✅1	1

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-09-04 03:07 UTC - **findings analyzed**: 51 - **repos with feedback**: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **48%** of reviewed PRs (20 / 42). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 1 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. | AC | repo | path:line | sev | problem | votes | resolved | seen | |---|---|---|---|---|---|---|---| | 1 | `gitea_admin/pragent` | `pilot/review/opencode_lens_config.py:58` | INFO | 🟡 [MEDIUM] _coerce_str is defined twice in this module — lines 29-30 and lines 58-59. The second de… | 👍0 👎0 | ✅1 | 1 | _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-09-05 03:14:32 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-09-05 03:14 UTC

• findings analyzed: 51
• repos with feedback: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 48% of reviewed PRs (20 / 42). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 1 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

AC	repo	path:line	sev	problem	votes	resolved	seen
1	gitea_admin/pragent	pilot/review/opencode_lens_config.py:58	INFO	🟡 [MEDIUM] _coerce_str is defined twice in this module — lines 29-30 and lines 58-59. The second de…	👍0 👎0	✅1	1

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-09-05 03:14 UTC - **findings analyzed**: 51 - **repos with feedback**: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **48%** of reviewed PRs (20 / 42). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 1 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. | AC | repo | path:line | sev | problem | votes | resolved | seen | |---|---|---|---|---|---|---|---| | 1 | `gitea_admin/pragent` | `pilot/review/opencode_lens_config.py:58` | INFO | 🟡 [MEDIUM] _coerce_str is defined twice in this module — lines 29-30 and lines 58-59. The second de… | 👍0 👎0 | ✅1 | 1 | _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

pragent-bot bot commented 2026-09-06 03:07:31 +00:00

Copy Link

Copy Source

pragent feedback report — 2026-09-06 03:07 UTC

• findings analyzed: 51
• repos with feedback: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site)
• case-review queue: 0 disagreement(s)

Restraint

⚠️ Bot posted findings on 48% of reviewed PRs (20 / 42). Above the 25% threshold — consider raising .pr-review.json:severity_threshold to medium or high for noisy repos, or adding patterns.deny to skip stylistic-only findings.

DoorDash rule (2026-07-06): excessive noise on clean code is its own failure mode. severity_threshold + patterns.deny are the knobs that dial restraint.

Top 0 false-positive candidates

Aggregated by posthash (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved.

none yet

Review each row before adding it to .pr-review.json:instructions. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting.

Top 1 accepted patterns

Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies.

AC	repo	path:line	sev	problem	votes	resolved	seen
1	gitea_admin/pragent	pilot/review/opencode_lens_config.py:58	INFO	🟡 [MEDIUM] _coerce_str is defined twice in this module — lines 29-30 and lines 58-59. The second de…	👍0 👎0	✅1	1

Promote widely-accepted patterns into the shared architecture.md on Nexus raw-hosted (or the per-repo additional_context_urls). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+.

Case-review queue (0)

No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here.

Where this report goes

• Per-repo actions (.pr-review.json:instructions, patterns.deny, severity_threshold): edit the file on main via a regular PR. The next PR review picks up the change automatically.
• Cross-repo actions (shared house-rules): update the PRAGENT_ADDITIONAL_CONTEXT_URL document on Nexus raw-hosted (canalhandia/architecture.md etc).
• Benchmark gate (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled posthash corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it.

Generated from /data/feedback.db by feedback_analyze.py.

# pragent feedback report — 2026-09-06 03:07 UTC - **findings analyzed**: 51 - **repos with feedback**: 4 (gitea_admin/pragent, netcracker/interview, techspark/suaspark-dashboard, techspark/suaspark-site) - **case-review queue**: 0 disagreement(s) ## Restraint ⚠️ Bot posted findings on **48%** of reviewed PRs (20 / 42). Above the 25% threshold — consider raising `.pr-review.json:severity_threshold` to `medium` or `high` for noisy repos, or adding `patterns.deny` to skip stylistic-only findings. > DoorDash rule (2026-07-06): *excessive noise on clean code is its own failure mode*. `severity_threshold` + `patterns.deny` are the knobs that dial restraint. ## Top 0 false-positive candidates Aggregated by `posthash` (path:line:severity:problem). Sort key = downvotes + unresolved + negation-phrase replies − upvotes − resolved. _none yet_ _Review each row before adding it to `.pr-review.json:instructions`. Human reactions are NOT ground truth (DoorDash, 2026-07-06: authors accept/reject for workflow reasons) — re-read the PR before acting._ ## Top 1 accepted patterns Aggregated by posthash. Sort key = upvotes + resolved − downvotes − unresolved − negation-phrase replies. | AC | repo | path:line | sev | problem | votes | resolved | seen | |---|---|---|---|---|---|---|---| | 1 | `gitea_admin/pragent` | `pilot/review/opencode_lens_config.py:58` | INFO | 🟡 [MEDIUM] _coerce_str is defined twice in this module — lines 29-30 and lines 58-59. The second de… | 👍0 👎0 | ✅1 | 1 | _Promote widely-accepted patterns into the shared `architecture.md` on Nexus raw-hosted (or the per-repo `additional_context_urls`). These become part of the prompt-cached prefix → ~0 marginal cost on step 2+._ ## Case-review queue (0) _No disagreements recorded yet. Once humans start reacting 👎 / leaving replies / not resolving bot comments, cases will appear here._ ## Where this report goes - **Per-repo actions** (`.pr-review.json:instructions`, `patterns.deny`, `severity_threshold`): edit the file on `main` via a regular PR. The next PR review picks up the change automatically. - **Cross-repo actions** (shared house-rules): update the `PRAGENT_ADDITIONAL_CONTEXT_URL` document on Nexus raw-hosted (`canalhandia/architecture.md` etc). - **Benchmark gate** (DoorDash pattern): before changing the model / prompt / context window, replay this report against the labeled `posthash` corpus. If a candidate addendum flips ≥ 1 currently-accepted finding into false-positive, drop it. _Generated from `/data/feedback.db` by `feedback_analyze.py`._

Sign in to join this conversation.

No Branch/Tag Specified

Branches Tags

main

feat/review-budget-governor

refactor/split-pilot-modules

test/eval-judges-trigger

refactor/pilot-architecture-langfuse

fix/dataset-item-url-safe-ids

feat/langfuse-evaluation

No results found.

Labels

Clear labels

AI-REVIEW

Trigger pragent-bot AI review

AI-USAGE

Opt-in: append AI token-usage summary + per-comment attribution to the review

No Label

Milestone

No items

No Milestone

Projects

Clear projects

No project

Assignees

Clear assignees

pragent-bot (pragent AI review bot) gitea_admin

No Assignees

1 Participants

Notifications

Subscribe

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: gitea_admin/pragent#8