Install
$ agentstack add skill-abysscn-oh-my-dag-retro ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
/retro — Engineering Retrospective
> Auto-generate an engineering retrospective from git history. Not from memory — from data.
Input
/retro → default: engineering retro over the last 7 days of git
/retro 24h → last 24 hours
/retro 14d → last 14 days
/retro 30d → last 30 days
/retro compare → this week vs last week comparison
Workflow
Step 1: Data collection (1 message, 3 parallel Bash)
# 1. Commits with stats
git log --since="{window_start}" --format="%H|%aN|%ai|%s" --shortstat
# 2. Per-file change frequency (hotspots)
git log --since="{window_start}" --name-only --format="" | sort | uniq -c | sort -rn | head -20
# 3. Commit timestamps for session detection
git log --since="{window_start}" --format="%at|%ai|%s" | sort -n
Step 2: Analysis (pure reasoning)
2a. Commit Type Mix
Classify by the type prefix of each commit message:
| Type | Meaning | Health indicator | |------|------|---------| | feat | new feature | primary output | | fix | fix | 40% + fix 50% = firefighting mode.
2b. Focus Score
focus = 1 - (unique_directories_touched / total_commits)
| Score | Meaning | |-------|------| | > 0.7 | highly concentrated (good: deep work) | | 0.4-0.7 | moderately spread (normal: multitasking) | | 2 hours → new session. Stats:
- Session count / average duration
- Peak hours (the densest commit window)
- Longest continuous work stretch
2d. Hotspots
File change frequency top 10. High-frequency files = potential instability zones or active development zones.
2e. Sprint Progress Check (if a progress map exists)
If your project keeps a sprint/progress map, read it → compare this cycle's plan vs actual completion.
Step 3: Output report
## Engineering Retro — {window} ({start} ~ {end})
### Summary
{1-2 sentences: the biggest achievement of this cycle}
Commits: {N} | Sessions: {M} | Focus: {score}
### Commit Mix
| Type | Count | % | Trend |
|------|-------|---|-------|
feat/fix/refactor/...
### Health Signals
- {building mode | firefighting mode | balanced}
- Fix share: {N}% {healthy | watch | warning}
- Test share: {N}% {sufficient | insufficient}
### Hotspots (Top 5 Most Changed Files)
| File | Changes | Type |
|------|---------|------|
### Session Patterns
- Peak hours: {HH:MM - HH:MM}
- Avg session: {N}h {M}min
- Longest session: {X}h
### Sprint Progress
- Planned: {items}
- Completed: {items} ({N}%)
- Blocked: {items}
### Retrospective
- **Keep**: {1-2 things that went well, anchored in data}
- **Improve**: {1-2 things to do differently, with specific suggestion}
- **Try**: {1 experiment for next sprint}
Step 4: Compare mode (/retro compare)
Compare two windows side by side:
| Metric | This Week | Last Week | Delta |
|--------|-----------|-----------|-------|
| Commits | ... | ... | +/- |
| feat% | ... | ... | +/- |
| fix% | ... | ... | +/- |
| Focus | ... | ... | +/- |
| Sessions | ... | ... | +/- |
Constraints
- Read-only — does not modify any file
- Output goes directly to the user, no file written
- Use
origin/branches for git queries (local main may be stale) - Use the user's local timezone for timestamps
- 0 commits in the window → report and suggest a different window
- Self-contained skill — does not read other docs
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: AbyssCN
- Source: AbyssCN/oh-my-dag
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.