docs(#239): add ADR-awareness steps to code-implementation skill

[fullsend-ai-coder]

Add two changes so the code agent checks Architecture Decision Records during planning:

1. Step 3 substep 1: mention checking for an ADR directory
   (docs/ADRs/, adr/, docs/architecture/decisions/) so the agent
   knows to consult them later in step 8.

2. Step 8: insert new substep 5 after "Follow cross-repo
   references" instructing the agent to search ADR files for
   keywords matching the area being changed. If an ADR constrains
   the approach, the agent must follow the ADR over ad-hoc issue
   recommendations. Renumber subsequent substeps (5-9 become
   6-10).

This prevents wasted code-review-fix cycles when an issue recommended approach conflicts with an accepted ADR.

---

Closes #239

Post-script verification

• Branch is not main/master (agent/239-adr-awareness-skill)
• Secret scan passed (gitleaks — c776c248ac9a91f10998ca7107e974a6c9d76514..HEAD)
• Pre-commit hooks passed (authoritative run on runner)
• Tests ran inside sandbox

[github-actions]

E2E tests did not run

E2E tests run automatically for org/repo members and collaborators on pull requests.

For other contributors, a maintainer must add the ok-to-test label after the latest push.

See E2E testing guide for details.

[fullsend-ai-review]

🤖 Finished Review · ❌ Failure · Started 1:14 PM UTC · Completed 1:14 PM UTC
Commit: c776c24 · View workflow run →

[guyoron1]

/fs-review

[fullsend-ai-review]

🤖 Finished Review · ✅ Success · Started 3:57 PM UTC · Completed 4:11 PM UTC
Commit: d8e3df7 · View workflow run →

[fullsend-ai-review]

Review — approve

PR: docs(#239): add ADR-awareness steps to code-implementation skill
Scope: Single file — internal/scaffold/fullsend-repo/skills/code-implementation/SKILL.md (22 additions, 9 deletions)
Linked issue: #239

Summary

Clean, well-scoped change that adds ADR-awareness to the code-implementation skill's planning phase. The PR implements exactly what issue #239 requested — no more, no less.

Dimensions reviewed

Dimension	Result
Correctness	✅ Step numbering is internally consistent (new substep 5 inserted, old 5–9 correctly renumbered to 6–10). Forward reference in step 3 ("step 8") accurately points to the "Plan the implementation" section. The ADR-precedence claim aligns with AGENTS.md's ADR immutability policy. No stale cross-references to old substep numbers found in the file or elsewhere in the repo.
Security	✅ No prompt injection, Unicode steganography, or bidirectional text overrides in the diff or PR body. The new ADR-reading instruction does not expand the trust boundary — ADR files are at the same trust level as CLAUDE.md and AGENTS.md, which the agent already reads.
Intent & coherence	✅ Change matches issue #239's authorization exactly. Commit type docs is correct per COMMITS.md. No scope creep — two changes (step 3 addition + step 8 new substep with renumbering) match the issue's proposed change verbatim.
Style & conventions	✅ New substep 5 in step 8 follows the existing pattern (N. **Bold title** — description). Markdown indentation is correct, including the 4-space continuation for the double-digit item 10.
Docs currency	✅ No external documents reference the old substep numbers. The only references to the code-implementation skill elsewhere are generic ("Follow the code-implementation skill for the step-by-step procedure") and unaffected by internal renumbering.
Cross-repo contracts	⏭ Skipped — no exported interfaces, schemas, or public APIs modified.

---

Labels: PR modifies agent skill documentation

[guyoron1]

/fs-fix

[fullsend-ai-coder]

🤖 Finished Fix · ✅ Success · Started 5:38 PM UTC · Completed 5:42 PM UTC
Commit: d8e3df7 · View workflow run →

[fullsend-ai-coder]

🔧 Fix agent — iteration 1 (human-triggered)

No changes made. The review agent approved PR #278 with no findings, and the human /fs-fix command included no instruction text. Nothing to fix.

Disagreed (1):

1. no actionable findings or instructions: The review agent approved this PR with no findings. The human triggered /fs-fix without providing an instruction (HUMAN_INSTRUCTION=none). There is nothing to fix — the working tree is clean and all review dimensions passed.

Tests: passed

_{Updated by fullsend fix agent}