Repository navigation
Make the Model Tiers Bullets Agree on What a Dispatch Naming No Model Runs At - #3538
Conversation
The highest bullet says a dispatch naming no model runs at its agent definition's model where one is set, and otherwise at the session's own. Two later bullets stated the session fallback without that condition. Delete the unconditional clause from the table-visibility bullet, and point the orchestrating-seat bullet at the highest bullet's statement. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Its rationale, "a choice nobody made", is false where an agent definition pins a model. The bullet's heading already states the rule, so the sentence goes rather than being qualified. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
A recorded local strict-review pass covers head |
|
Important
This repository does not receive automatic reviews because it has fewer than 10 stars.
|
There was a problem hiding this comment.
🟢 Approval recommended
All reviewed changes are documentation updates with no unresolved review findings.
No open issues 🎉
What changed in this PR
This documentation-only pull request removes contradictory model fallback wording from the canonical Model Tiers guidance and regenerates the affected skill distributions and source digests.
Changes:
- Removes the conflicting session-model fallback wording from
AGENTS.md. - Regenerates the authored, GitHub, and Claude skill copies.
- Updates the corresponding source-digest files.
| File | Description |
|---|---|
AGENTS.md |
Updates the canonical Model Tiers guidance. |
.github/skills/unattended-handoff/SKILL.md |
Regenerated GitHub skill copy. |
.github/skills/local-strict-review/SKILL.md |
Regenerated GitHub skill copy. |
.github/skills/backlog-burndown/SKILL.md |
Regenerated GitHub skill copy. |
.claude-plugin/fleet-skills/skills/unattended-handoff/SKILL.md |
Regenerated Claude skill copy. |
.claude-plugin/fleet-skills/skills/local-strict-review/SKILL.md |
Regenerated Claude skill copy. |
.claude-plugin/fleet-skills/skills/backlog-burndown/SKILL.md |
Regenerated Claude skill copy. |
.claude-plugin/fleet-skills/.source-digests/unattended-handoff |
Updated source digest. |
.claude-plugin/fleet-skills/.source-digests/local-strict-review |
Updated source digest. |
.claude-plugin/fleet-skills/.source-digests/backlog-burndown |
Updated source digest. |
.agents/skills/unattended-handoff/SKILL.md |
Updated authored skill include. |
.agents/skills/local-strict-review/SKILL.md |
Updated authored skill include. |
.agents/skills/backlog-burndown/SKILL.md |
Updated authored skill include. |
Review effort: Lite
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## develop #3538 +/- ##
========================================
Coverage 81.58% 81.58%
========================================
Files 26 26
Lines 8972 8972
========================================
Hits 7320 7320
Misses 1652 1652
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:
|
…y Tests, and Copilot Transient Fixes (#3544) ## Summary Promotes `develop` to `main`, carrying round 1 of the resync-issue burndown: - #3538: `AGENTS.md` "Model Tiers" now states the no-model dispatch result once, conditionally, in the `highest` bullet. The two bullets that contradicted it lost their unconditional claims. The three skills including the section are regenerated. - #3539: `GOVERNANCE.md` "Durable Knowledge and Self-Improvement" describes the standing-issue handoff lifecycle, matching `AGENTS.md` "Session Scope". `agent-conduct` is regenerated. - #3540: `spec/files.json` gains a `repoOwnedTrees` list declaring `tests`, and `spec/audit.py` `hub_only_paths` leaves those paths out. A carrier's own `tests/` file is no longer reported as hub-only. `spec/validate.py` shape-checks the list, and `AUDIT.md`, `RESYNC.md` and the project-types assertion state the exclusion. - #3542: `pr_review.py wait` and `status` read the Copilot run log before calling an error refusal quota. A `too many files changed` line is transient and earns one re-request, as `COPILOT_ERROR_TRANSIENT` on the existing exit 41. - #3545: promotion fix for Copilot's finding on this pull request. `spec/validate.py` refuses a `repoOwnedTrees` source that overlaps a carried `trees[].source` or holds a `baseline` path. Fixes #3529 Fixes #3504 Fixes #3528 Fixes #3510 🤖 Generated with [Claude Code](https://claude.com/claude-code)
Summary
AGENTS.md"Model Tiers" gave two answers for a dispatch that names no model. Thehighestbullet states the conditional form: the agent definition's model where one is set, otherwise the session's own. Two later bullets stated only the session fallback.The
highestbullet stays the single statement of the result.scripts/build_dist.pyregenerated the include regions inbacklog-burndown,local-strict-review, andunattended-handoff, plus their generated copies.Pre-existing findings, filed
AGENTS.md"Model Tiers", the paragraph under the tier table, and thespec/model-tiers.jsonnote: an unconditional "session's own model" claim for a tool with no row. Classpre-existing, filed as Make the Model Tiers Tool-Fallback Sentence and the Spec Note Agree With the No-Model Rule #3537, since the two change together and the spec file is outside this change.Closes on promotion: #3529
🤖 Generated with Claude Code