Repository navigation
Conversation
## Summary
Commits the hub's own rendered agent policy files, the first part of the
policy-driven review pilot. `scripts/agent_policy.py write` rendered
them from `registry/repos.json`. ProjectTemplate takes the `release`
workflow model and has no `agentPolicy` override. The hub then carries
its rendered files before any account-level automatic review toggle goes
off.
- `.github/agent-policy.json` is new. It is the hub default: Copilot,
CodeRabbit, and Qodo each `trigger: auto`, and `guard: {}` for the
release model.
- `.pr_agent.toml` gains `[config] disable_auto_feedback = false`,
stating Qodo's automatic review explicitly. It matches pr-agent's own
default.
- `.coderabbit.yaml` renders byte-identical and is unchanged.
Behavior does not change. `scripts/pr_review.py` reads the same bot
settings it reads today from `spec/agent-policy.default.json`, and
`gh-write-guard` reads `guard: {}` as no exemption, the same as an
absent file. Moving the hub to explicit review is a later change,
through a registry `agentPolicy` override.
The files were written with `agent_policy.py write` rather than
`repo-config/configure.sh apply --checkout`, since `apply` also writes
the repository's settings, labels, and rulesets. The files are the same
either way.
## Verification
- `agent_policy.py check --ref feature/render-agent-policy-pilot`
reports all four items `ok`: the policy file equals the render,
`.coderabbit.yaml`'s `auto_review.enabled` and `base_branches`, and
`.pr_agent.toml`'s `disable_auto_feedback`.
- A recorded local strict pass at the `frontier` tier found 0 findings,
and `prose_lint.py --diff` printed no violation.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
|
Important
This repository does not receive automatic reviews because it has fewer than 10 stars.
|
There was a problem hiding this comment.
🟢 Approval recommended
No active review issues were identified in the policy-file additions or the Qodo feedback configuration change.
No open issues 🎉
What changed in this PR
Promotes the hub’s registry-rendered agent policy files to main, preserving the existing automatic review behavior for Copilot, CodeRabbit, and Qodo.
Changes:
- Adds
.github/agent-policy.jsonwith the rendered review policy and empty release-model guard. - Enables Qodo automatic feedback in
.pr_agent.toml. - No actionable findings were identified.
| File | Description |
|---|---|
.pr_agent.toml |
Enables Qodo automatic feedback explicitly. |
.github/agent-policy.json |
Adds the rendered agent review policy and empty release-model guard. |
Review effort: Lite
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #3898 +/- ##
=======================================
Coverage 77.87% 77.87%
=======================================
Files 34 34
Lines 12085 12085
=======================================
Hits 9411 9411
Misses 2674 2674
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
Summary
Promotes #3895 to
main..github/agent-policy.jsonand adds[config] disable_auto_feedback = falseto.pr_agent.toml, both rendered fromregistry/repos.jsonbyscripts/agent_policy.py. The render matches today's behavior: Copilot, CodeRabbit, and Qodo each review automatically, andguardis empty for the release model. This is the first part of the policy-driven review pilot. Once it is onmain,configure.sh checkstops reporting agent-policy drift on ProjectTemplate. A later registryagentPolicyoverride moves the hub to explicit review.The feature pull request carried a recorded local strict pass with no findings, and Copilot's round on it was clean.
🤖 Generated with Claude Code