Add aidd-fix skill for structured bug fixing workflow#103
Conversation
- Add ai/skills/fix/SKILL.md: structured 6-step TDD fix process - Add ai/commands/aidd-fix.md: command entry point for the skill - Add 🐛 /aidd-fix to please.mdc Commands section - Add task index mapping to AGENTS.md (fix bug => /aidd-fix) https://claude.ai/code/session_01ESPcHDaPRBLaTxDspPMd3Z
There was a problem hiding this comment.
Pull request overview
This pull request introduces a new aidd-fix skill that standardizes the bug-fixing workflow by enforcing test-driven development (TDD) practices through a structured 6-step process. The skill ensures quality through mandatory test writing before implementation and e2e validation before commits, addressing common pitfalls in bug-fixing workflows.
Changes:
- Adds comprehensive aidd-fix skill documentation with a 6-step AIDD fix process (context validation, epic documentation, TDD test writing, implementation, e2e testing, and commit)
- Registers the new
/aidd-fixcommand in the command system and available commands list - Adds task index mapping for bug-fixing discoverability
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated no comments.
| File | Description |
|---|---|
| ai/skills/fix/SKILL.md | New skill file defining the complete 6-step AIDD fix process with constraints, requirements format, and test-driven workflow |
| ai/commands/aidd-fix.md | Command entry point that loads and executes the fix skill |
| ai/rules/please.mdc | Registers the new 🐛 /aidd-fix command in the available commands list |
| AGENTS.md | Adds task index mapping "fix bug => /aidd-fix" for command discoverability |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
|
@cursoragent please /review |
|
Now I'll conduct a thorough code review following the review.mdc guidelines. 🔬 Code Review - PR #97: Add aidd-fix skill for structured bug fixing workflowOverviewThis PR introduces a new 🎯 Review AnalysisFiles Changed:
|
| Criterion | Status | Notes |
|---|---|---|
| Documentation clarity | ✅ | Clear, well-structured steps |
| Pattern consistency | Minor inconsistencies in file references | |
| Completeness | ✅ | All workflow steps covered |
| TDD discipline | ✅ | Properly enforces test-first |
| Error prevention | ✅ | Explicit constraints included |
| Discoverability | Requires index generation |
🏗️ Architecture Review
✅ Separation of concerns: Command file cleanly delegates to skill file
✅ Integration: Properly references existing tools (tdd.mdc, task-creator.mdc, review.mdc)
✅ Workflow design: Systematic, prevents skipping steps
✅ Traceability: Links requirements → epics → tests → code
📋 Actionable Recommendations
Required Before Merge:
-
Fix file reference in
ai/skills/fix/SKILL.md:71- Using `review.mdc`: + Using `/review`:
-
Verify index generation - Run pre-commit hook or manually verify:
node bin/aidd.js --index git add ai/**/index.md
Suggested Improvements (Optional):
-
Add usage example to SKILL.md showing a minimal end-to-end bug fix
-
Standardize references throughout SKILL.md to use command names (
/review,/commit) instead of file paths -
Consider adding a brief "When to use" section to help agents decide when to invoke this skill vs. handling bugs directly
📝 Test Plan Review
Stated: "N/A — This is documentation and command registration."
Assessment: ✅ Appropriate. The skill will be validated through actual usage in bug-fixing workflows. No automated tests needed for documentation.
✨ Final Recommendation
Status:
This PR provides valuable standardization for bug-fixing workflows and follows good patterns. The implementation is sound, but requires:
- Fixing the incorrect file reference on line 71
- Ensuring index files are generated before merge
Once these are addressed, the PR is ready to merge.
Risk Level: Low (documentation only, no runtime changes)
💬 Summary
This skill introduces a well-structured, systematic approach to bug fixing that enforces TDD discipline and prevents common mistakes. The step-by-step workflow with explicit constraints will help maintain code quality. Minor fixes needed before merge, but overall this is a valuable addition to the AIDD framework.
Replace file path references with command names throughout: - task-creator.mdc → /task - tdd.mdc → /task - review.mdc → /review (incorrect file ref) - commit.md → /commit https://claude.ai/code/session_01ESPcHDaPRBLaTxDspPMd3Z
|
@cursoragent please review fixes in 5f8744e |
🔬 Review of Fixes - Commit 5f8744eSummary of ChangesThe commit attempted to standardize command references by replacing file path references with command names. Here's what changed: Changes Made:
|
| Issue | Previous Status | Current Status | Notes |
|---|---|---|---|
| Missing index files | ✅ RESOLVED | Generated but need to commit | |
Incorrect review.mdc reference |
✅ FIXED | Line 70 now correctly uses /review |
|
| Mixed reference styles | Improved but introduced new error on line 45 |
🚨 Remaining Issues
Critical Issue: Incorrect Step 3 Reference
The fix commit over-corrected and introduced a semantic error:
- Before: Referenced
tdd.mdc(a rule file, correct context) - After: References
/task(wrong command for the context)
Impact: Agents following this workflow will invoke task planning when they should be applying TDD principles for test writing.
📋 Required Actions Before Merge
-
Fix Step 3 reference in
ai/skills/fix/SKILL.md:45- Either reference
tdd.mdc(rule file) - Or remove the reference entirely
- Do NOT reference
/task
- Either reference
-
Commit the generated index files:
git add ai/index.md ai/commands/index.md ai/skills/index.md ai/skills/fix/index.md git commit -m "chore: generate index files for skills directory"
✨ Overall Assessment
Progress: The fix commit resolved the main issue (line 70) but introduced a new semantic error (line 45).
Recommendation:
The PR is very close to being ready. Once the Step 3 reference is corrected and index files are committed, this will be good to merge.
Next Steps:
- Fix line 45 reference issue
- Commit generated index files
- Final verification, then approve for merge
…mands block
- Drop "read please.mdc" constraint (already in every invoked command)
- Drop Step 3 item 4 (duplicates top-level "Never implement before writing a failing test")
- Drop "using /review" from Step 5 item 2 (redundant with section header)
- Remove Commands {} block (belongs in please.mdc, not the skill itself)
https://claude.ai/code/session_01ESPcHDaPRBLaTxDspPMd3Z
There was a problem hiding this comment.
Pull request overview
Copilot reviewed 4 out of 4 changed files in this pull request and generated 5 comments.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
🔬 Agent Skills Spec Cross-ReferenceCross-referencing 🚨
|
There was a problem hiding this comment.
Pull request overview
Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
/review is review-only and must not imply executing commands. Move `npm run test:e2e` to a standalone instruction and make /review an optional follow-up for self-review. https://claude.ai/code/session_01ESPcHDaPRBLaTxDspPMd3Z
Adds missing ai/skills/index.md and ai/skills/fix/index.md so the aidd-fix skill is discoverable via the index system. Also updates ai/index.md and ai/commands/index.md to reflect current directory state. https://claude.ai/code/session_01ESPcHDaPRBLaTxDspPMd3Z
New skills from main (aidd-ecs, aidd-layout, aidd-lit, aidd-namespace, aidd-observe, aidd-react, aidd-service, aidd-structure) needed indexing. https://claude.ai/code/session_01ESPcHDaPRBLaTxDspPMd3Z
…IndexRecursive declaration
… declarations - Switch tsconfig to NodeNext module resolution (supports import attributes, import.meta.url) - Add paths for aidd/utils and aidd/server package self-imports in exports.test.js - Install @types/node and @types/fs-extra - Add d.ts sidecars for utils (async-pipe, pipe, compose) and lib/release-helpers so TypeScript uses declarations instead of checking .js implementations - Add // @ts-check to all test files (per-file opt-in; avoids cascading checkJs into src/server implementation files) - Fix implicit-any issues: let tempDir initializers, consoleSpy JSDoc type, catch-block error casts, optional chaining for .find() and parseFrontmatter results
Tests now verify type correctness end-to-end. 88 type errors in src/server/ remain — tracked in the epic, to be fixed in follow-up tasks.
- Add types/error-causes.d.ts ambient declaration (temporary local fix; upstream tracked at github.com//issues/109) - Wire types/ into tsconfig include - Annotate all implicit-any parameters in src/server/ with JSDoc @param and inline /** @type */ casts on destructured context arguments - Fix headers nullability in with-auth.js getSession call - Fix response.locals optional-chaining via any cast in with-auth.js - Fix Record<> index signature in with-config.js loadConfigFromEnv - Fix asyncPipe spread cast and sanitizeHeaders type in create-route.js - Fix createServer mock return cast in test-utils.js
Co-authored-by: Eric Elliott <support@paralleldrive.com>
Description
Introduces a new
aidd-fixskill that provides a structured, step-by-step process for fixing bugs and implementing review feedback. This skill enforces test-driven development (TDD) practices and ensures quality through mandatory e2e testing before commits.Changes
ai/skills/fix/SKILL.md (new): Comprehensive skill documentation defining the 6-step AIDD fix process:
ai/commands/aidd-fix.md (new): Command entry point that loads and executes the fix skill
AGENTS.md: Added task index mapping
fix bug => /aidd-fixfor discoverabilityai/rules/please.mdc: Registered the new
🐛 /aidd-fixcommand in the available commands listRationale
This skill standardizes the bug-fixing workflow across the team by:
Test Plan
N/A — This is documentation and command registration. The skill will be validated through actual usage in bug-fixing workflows.
https://claude.ai/code/session_01ESPcHDaPRBLaTxDspPMd3Z
Note
Medium Risk
Touches installer/CLI behavior and TypeScript config (NodeNext +
checkJs), which can affect installs and downstream typechecking despite being mostly additive and well-tested.Overview
Adds a new bug-fixing workflow command
/aidd-fixand supporting skill docs, wiring it intoplease.mdc,AGENTS.md, and generatedai/**/index.mdnavigation.Introduces an
aidd-custom/customization layer (docs + defaultconfig.yml+ auto-generated index) and updates the installer (executeClone) and pre-commit hook to create/index it, while ensuringconfig.ymlis never overwritten (even with--force) and dry-run output reports its status.Refactors
AGENTS.mddirective management to include the newaidd-customdirective, avoid duplicating the auto-appended wrapper on upgrades, and aligns exports/types (camelCase constants, updated.d.ts). Addslib/cli-core.d.ts, normalizesgenerateAllIndexeserror shape, extends index generation to optionally indexaidd-custom/, and turns on repo-wide JS typechecking (checkJs, NodeNext resolution) with new ambient/types and updated tests.Written by Cursor Bugbot for commit eebf4ed. This will update automatically on new commits. Configure here.