chore: update agent prompts to read AGENTS.md and CLAUDE.md

- Added "Read `AGENTS.md` and `CLAUDE.md`" instruction to the Preflight/Startup/Checklist sections of 8 agent prompts:
  - src/prompts/daily/audit-agent.md
  - src/prompts/daily/const-refactor-agent.md
  - src/prompts/daily/content-audit-agent.md
  - src/prompts/daily/decompose-agent.md
  - src/prompts/daily/deps-security-agent.md
  - src/prompts/daily/docs-code-investigator.md
  - src/prompts/daily/innerhtml-migration-agent.md
  - src/prompts/daily/perf-agent.md
  - src/prompts/daily/protocol-research-agent.md
  - src/prompts/daily/test-audit-agent.md
  - src/prompts/daily/torch-garbage-collection-agent.md
  - src/prompts/weekly/feature-proposer-agent.md

This ensures all agents are aware of repository policies, including KNOWN_ISSUES.md and src/issues usage.

Co-authored-by: PR0M3TH3AN <[email protected]>
This commit is contained in:
google-labs-jules[bot]andPR0M3TH3AN committed 2026-02-14 06:32:19 +00:00
1 parent 9891e9e632
commit f50c31c884
12 files changed
+19 -7

No files matched your search

+1
View File
@@ -33,6 +33,7 @@ PRIMARY WORKFLOW (end-to-end)
1. **Preflight**
- Ensure you are on `<default-branch>`:
- `git checkout <default-branch> && git pull --ff-only --ff-only`
- Read `AGENTS.md` and `CLAUDE.md`.
- Record environment:
- `node -v`, `npm -v` (or `pnpm`/`yarn`), OS, current `git rev-parse HEAD`
- Create the artifact directory: `artifacts/audit/$(date +%F)/`
@@ -43,6 +43,7 @@ DETAILED WORKFLOW (step-by-step)
git checkout <default-branch>
git pull --ff-only
```
- Read `AGENTS.md` and `CLAUDE.md`.
- Record baseline:
- `git rev-parse HEAD` (commit SHA)
- `node -v`, `npm -v`
+3 -2
View File
@@ -242,8 +242,9 @@ QUALITY BAR & BEHAVIOR
===============================================================================
FIRST-RUN CHECKLIST (practical)
1. Create files in `context/`, `todo/`, `decisions/`, `test_logs/`.
2. Produce inventory: `artifacts/docs-audit/YYYY-MM-DD/inventory.md`.
1. Read `AGENTS.md` and `CLAUDE.md`.
2. Create files in `context/`, `todo/`, `decisions/`, `test_logs/`.
3. Produce inventory: `artifacts/docs-audit/YYYY-MM-DD/inventory.md`.
3. Run verification for highest-priority pages (upload/contribute).
4. Update `/content` pages and prepare diff.
5. Validate end-to-end for changed pages and capture artifacts.
+1
View File
@@ -244,6 +244,7 @@ FIRST-RUN CHECKLIST (execute now)
1. Checkout <default-branch> & record baseline:
- Read `AGENTS.md` and `CLAUDE.md`.
```
git checkout <default-branch>
git pull --ff-only
+3 -2
View File
@@ -48,8 +48,9 @@ Record chosen manager and versions (`node -v`, `npm -v` / `pnpm -v` / `yarn -v`)
DAILY WORKFLOW (run every day / scheduled)
A. Bootstrap & baseline
1. `git checkout <default-branch> && git pull --ff-only`
2. Record environment: `node -v` and package-manager version(s).
3. Ensure a clean workspace and no local changes.
2. Read `AGENTS.md` and `CLAUDE.md`.
3. Record environment: `node -v` and package-manager version(s).
4. Ensure a clean workspace and no local changes.
B. Clean install
- npm: `npm ci`
+2 -1
View File
@@ -202,7 +202,8 @@ EXAMPLE OUTPUT (minimal)
===============================================================================
FIRST-RUN CHECKLIST
1. `git checkout <default-branch> && git pull --ff-only --ff-only`
2. Produce candidate list:
2. Read `AGENTS.md` and `CLAUDE.md`.
3. Produce candidate list:
```
git ls-files 'js/**/*.js' | xargs -n1 wc -l | sort -rn | head -n 20
@@ -28,6 +28,7 @@ Commit these artifacts in the PR branch so reviewers can reproduce.
WORKFLOW — top-level steps (one file only)
1. **Preflight**
- `git checkout <default-branch> && git pull --ff-only --ff-only`
- Read `AGENTS.md` and `CLAUDE.md`.
- `node -v` and `npm -v` recorded in `context/CONTEXT_<timestamp>.md`
- Run the innerHTML report:
```
+1
View File
@@ -50,6 +50,7 @@ DAILY WORKFLOW (run each day)
1. **Preflight**
- Ensure default branch.
- Pull latest and confirm artifacts exist.
- Read `AGENTS.md` and `CLAUDE.md`.
- Record environment: Node, package manager, OS, browser versions if relevant.
2. **Search & Inventory**
+3 -2
View File
@@ -21,8 +21,9 @@ Mission: identify external protocol/spec dependencies used by this codebase, map
## Runbook
1. Scan the repo for protocol/spec references and external formats.
2. Update `PROTOCOL_INVENTORY.md` with concrete code pointers.
1. Read `AGENTS.md` and `CLAUDE.md`.
2. Scan the repo for protocol/spec references and external formats.
3. Update `PROTOCOL_INVENTORY.md` with concrete code pointers.
3. Validate behavior with targeted unit/integration tests.
4. Classify gaps and propose minimal remediations.
5. Publish a dated report: `protocol-report-YYYY-MM-DD.md`.
+1
View File
@@ -43,6 +43,7 @@ Always read `AGENTS.md` and `docs/KNOWN_ISSUES.md` for repo-specific caveats (lo
DAILY WORKFLOW (run every day or as scheduled)
1) **Discover the test runner**
- Read `AGENTS.md` and `CLAUDE.md`.
- Inspect `package.json`:
- `test` script (e.g., `jest`, `vitest`, `mocha`, `cypress`, `playwright`).
- Test-related devDependencies.
@@ -37,6 +37,7 @@ Out of scope:
1. Confirm repo root:
- `pwd`
- Read `AGENTS.md` and `CLAUDE.md`.
- Must resolve to `/workspace/TORCH` (or a child directory).
2. Generate candidate list:
@@ -22,6 +22,7 @@ Out of scope:
## Weekly workflow
1. **Analyze** the repository context.
- Read `AGENTS.md` and `CLAUDE.md`.
- Read `README.md`, `dashboard/`, `landing/`, and `src/` to understand the project's purpose and structure.
- Look for gaps in tooling, documentation, or user experience.
2. **Brainstorm** a list of potential features.