feat(reflection): add skill for systematic post-execution evaluation - #203
Conversation
Adds reflection skill following TDD methodology (RED-GREEN-REFACTOR): - RED: Documented baseline behavior without skill (3 scenarios) - GREEN: Wrote minimal skill addressing observed issues - REFACTOR: Added rationalization table and red flags Features: - 4-step evaluation process (CAPTURE-EVALUATE-DIAGNOSE-RECOMMEND) - Checklist-based evaluation (no subjective metrics) - Recursive "why?" questioning (min 3x) for root cause - Rationalization detection with 12 common excuses - Red flags for pressure scenarios - Problem classification (skill bug/usage/environment/edge case) Tested with: - Rule violation scenarios (clear failure detection) - Success cases (appropriate recognition) - Efficiency problems (technical success vs optimal) - Pressure scenarios (rationalization resistance) Word count: 991 (within acceptable range for comprehensive skill) Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
📝 WalkthroughWalkthroughA new documentation file defines a Reflection workflow for post-skill evaluation, including structured steps for capturing execution details, evaluating outcomes, diagnosing root causes through iterative questioning, and recommending improvements. The document standardizes the reflection output format as a Markdown report. Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~12 minutes Poem
🚥 Pre-merge checks | ✅ 3✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Fix all issues with AI agents
In `@plugins/me/skills/reflection/SKILL.md`:
- Line 222: Update the table cell in SKILL.md for the row with "Already spent
enough time" to use a standard unit spacing by replacing "5min" with "5 min" (or
"5 minutes") so the entry reads e.g. | "Already spent enough time" | 5 min saves
hours later. Sunk cost fallacy |; locate the exact string "Already spent enough
time" to make the correction.
🧹 Nitpick comments (3)
plugins/me/skills/reflection/SKILL.md (3)
28-33: Add language identifiers to fenced code blocks.markdownlint flags these fences without a language; use
textfor diagrams and lists.Proposed fix
-``` +```text 1. CAPTURE (30s): Identify last skill/command from conversation 2. EVALUATE (1min): Checklist-based assessment 3. DIAGNOSE (2min if issues): Recursive "why?" to root cause 4. RECOMMEND (1min): Specific, actionable improvements@@
-+text
Observe symptom
↓
"Why?" → 1st cause
↓
"Why?" → 2nd cause
↓
"Why?" → Root cause@@ -``` +```text Symptom: create-pr skill didn't check merge conflicts ↓ Why? → Didn't run git merge-tree command ↓ Why? → Skipped that step ↓ Why? → Rationalized "already checked locally" ↓ Root cause: Skill's rationalization table doesn't cover this case</details> Also applies to: 96-104, 110-120 --- `64-79`: **Surround headings with blank lines (markdownlint MD022).** Add a blank line before these subheadings to satisfy markdownlint. <details> <summary>Example fix (apply similarly to all flagged headings)</summary> ```diff ### Task Completion - [ ] Achieved intended goal? - [ ] Performed all required steps? - [ ] Reached expected final state? + ### Skill Rule Compliance - [ ] Followed all MUST/ALWAYS rules? - [ ] Violated no NEVER rules? - [ ] Followed recommended workflow?Also applies to: 237-250
14-25: Add explicit activation guidelines for the skill.The file has “When to Use,” but the SKILL.md structure guideline calls for context-aware activation guidelines; add an explicit “Activation Guidelines” section with triggers/prerequisites.
Based on learnings: “Applies to **/skills//SKILL.md : Create SKILL.md files in skills//SKILL.md structure with context-aware activation guidelines and domain-specific best practices.”
| |--------|---------| | ||
| | "Technically successful" | Rule violation = failure | | ||
| | "Small violation, ignore" | Small violation is still violation | | ||
| | "Already spent enough time" | 5min saves hours later. Sunk cost fallacy | |
There was a problem hiding this comment.
Fix “5min” spelling.
Use a standard unit spacing (e.g., “5 min” or “5 minutes”).
Proposed fix
-| "Already spent enough time" | 5min saves hours later. Sunk cost fallacy |
+| "Already spent enough time" | 5 min saves hours later. Sunk cost fallacy |📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| | "Already spent enough time" | 5min saves hours later. Sunk cost fallacy | | |
| | "Already spent enough time" | 5 min saves hours later. Sunk cost fallacy | |
🧰 Tools
🪛 LanguageTool
[grammar] ~222-~222: Ensure spelling is correct
Context: ...ation | | "Already spent enough time" | 5min saves hours later. Sunk cost fallacy | ...
(QB_NEW_EN_ORTHOGRAPHY_ERROR_IDS_1)
🤖 Prompt for AI Agents
In `@plugins/me/skills/reflection/SKILL.md` at line 222, Update the table cell in
SKILL.md for the row with "Already spent enough time" to use a standard unit
spacing by replacing "5min" with "5 min" (or "5 minutes") so the entry reads
e.g. | "Already spent enough time" | 5 min saves hours later. Sunk cost fallacy
|; locate the exact string "Already spent enough time" to make the correction.
1 participant
Summary
Adds reflection skill for systematic post-execution evaluation of skills/commands.
Key Features
TDD Methodology
Usage
```bash
/reflection # Evaluates last skill automatically
/reflection create-pr # Evaluates specific skill
```
Test Results
Implementation
plugins/me/skills/reflection/SKILL.md🤖 Generated with Claude Code
Summary by CodeRabbit
✏️ Tip: You can customize this high-level summary in your review settings.