timeline
Skill12 starsChanged 8 months agoMITView on GitHub
Group the current branch's commit history into semantic phases along a timeline, and generate two review documents: architecture evolution + code review.
5 other projects use a version of this.
Adopted by
Most recently taken up by
- chellehoniker/claude-code-storytellerostheir own version, first seen Oct 20, 2026
- anthonysuherli/br8ntheir own version, first seen Oct 13, 2026
- hailneed/agent-blackboxtheir own version, first seen Aug 19, 2026
- AlexeyPerov/Unity-Open-MCPtheir own version, first seen Jul 14, 2026
- npow/claude-skillstheir own version, first seen May 8, 2026
Other versions5 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 5
- AlexeyPerov/Unity-Open-MCP★ 17same name
- npow/claude-skills★ 4same name
- anthonysuherli/br8n★ 1same name
- chellehoniker/claude-code-storytelleros★ 1same name
- hailneed/agent-blackbox★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository29 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- add-testsSkill
- best-practiceSkill
- bootstrapSkill
- briefSkill
- changesSkill
- claude-code-reviewSkill
- claude-plan-reviewSkill
- codex-code-reviewSkill
- codex-plan-reviewSkill
- commit-pushSkill
- depsSkill
- design-researchSkill
- designSkill
- doc-driftSkill
- doc-syncSkill
- doc-tidySkill
- executeSkill
- explainSkill
- full-briefSkill
- goSkill
- grandmaSkill
- planSkill
- post-design-reviewSkill
- recordSkill
- self-reviewSkill
- tech-researchSkill
- testSkill
The file
We split this skill into the rules it gives an agent
8 rules. 0 are common, 7 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
7
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 8
No arguments: compare against the default base branch (develop if it exists, else main, else master)
- Distinctive: line 24
Phase markers in commit messages: if a commit self-annotates with a Phase, respect that first
- Distinctive: line 36
Read the full content of key changed files to understand context — not just diff fragments.
- Distinctive: line 86
Use green to highlight parts newly added in this phase, gray for existing context
- Distinctive: line 87
If this phase modifies a structure introduced in a previous phase, use orange to highlight it
- Distinctive: line 171
If a later phase modifies code from an earlier phase, annotate it in that later phase as "modifies Phase X's ...".
- Distinctive: line 233
Make sure the project compiles, then run /qq:test to verify ↓ Optional: run /qq:brief to generate final-state docs for last-pass coverage ↓ …
- Very specific: line 252
Write three files to Docs/qq/<branch-name>/:
“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Respond in the user's preferred language (detect from their recent messages, or fall back to the language setting in CLAUDE.md).
Arguments: $ARGUMENTS
- No arguments: compare against the default base branch (develop if it exists, else main, else master)
--base <branch>: specify a custom base branch for comparison
Execution Steps
1. Collect Commit History
git log <base>..HEAD --oneline --reverse --format="%h %ai %s"2. Group Commits into Semantic Phases
Grouping criteria (in order of priority):
- Semantic affinity: consecutive commits on the same feature/subsystem go in the same group
- Natural breakpoints: merge commits, date gaps > 1 day, module switches
- Phase markers in commit messages: if a commit self-annotates with a Phase, respect that first
Give each Phase a semantic name (e.g. "Player Health System", not "Phase 1"). Target 5–10 Phases (too few loses the timeline's value; too many becomes per-commit annotation); a Phase of more than 15 commits may be split into sub-phases. Grouping is the judgment the whole document rests on — settle it before writing.
3. Analyze Changes per Phase
For each Phase:
git diff <phase_first_commit>~1..<phase_last_commit> --stat
git diff <phase_first_commit>~1..<phase_last_commit> -- '*.cs'Read the full content of key changed files to understand context — not just diff fragments.
4. Generate Document A: Architecture Evolution Timeline
Format:
# Architecture Evolution Timeline
> X phases, Y commits, Z-day development span
> Branch: `<branch>`, base: `<base>`
---
## Phase 1: <Semantic Name> (<date range>, N commits)
> One-sentence summary
<details>
<summary>Commits</summary>
- `hash1` message1
- `hash2` message2
</details>
### Architecture Changes
#### [Tier 1] Change Title (if any)
**Scope**: ...
**Nature of change**: ...
\```mermaid
<diagram of this phase's changes>
\```
#### [Tier 2] Change Title (if any)
...
### Dependencies Introduced This Phase
- `ModuleA` → `ModuleB` (new reference)
### Cumulative State
> Overall progress to this point: X completed, next phase will Y
---
## Phase 2: <Semantic Name> ...Diagram requirements:
- Each diagram shows only what this phase changed, not the accumulated state — a reviewer should see "what was added in this phase"
- Use green to highlight parts newly added in this phase, gray for existing context
- If this phase modifies a structure introduced in a previous phase, use orange to highlight it
- A Phase with no architecture changes (pure bug fix/UI) gets a one-line note instead of a diagram
5. Generate Document B: Code Review Timeline
Format:
# Code Review Timeline
> X phases, Y commits
> Branch: `<branch>`, base: `<base>`
## Review Priority Quick Reference
| Phase | P0 | P1 | P2 | Files | Est. Time | Core Risk |
|-------|----|----|----|-------|-----------|-----------|
| 1. Name | 0 | 2 | 1 | 5 | 10 min | No major risk |
| 2. Name | 3 | 1 | 0 | 12 | 25 min | Global static state isolation |
| ... | | | | | | |
Time estimation rules:
- Each P0 item ~5 min (requires reading context + verification)
- Each P1 item ~2 min (quick check)
- Each P2 item ~0.5 min (quick scan)
- Round up to the nearest 5 minutes
**Recommended review order**: sort by P0 count descending — review the highest-risk phases when most focused
---
## Phase 1: <Semantic Name> (<date range>, N commits)
> One-sentence summary
<details>
<summary>Commits</summary>
- `hash1` message1
- `hash2` message2
</details>
### Files to Review
C# files touched in this phase, by priority:
| File | Priority | Change Summary |
|------|----------|---------------|
| `Assets/Scripts/.../GameManager.cs` | P0 | Core game loop changes |
| `Assets/Scripts/.../PlayerController.cs` | P0 | Input handling refactor |
| `Assets/Scripts/.../InventorySystem.cs` | P1 | Data model migration |
| ... | P2 | ... |
File list generation rules:
- Obtain from `git diff <phase_first_commit>~1..<phase_last_commit> --name-only -- '*.cs'`
- Exclude pure asset/config files, and test files under `Tests/` (unless the tests themselves have P0/P1 review items)
- Exclude pure Editor tool files (unless they have review items)
- Each file is annotated with its highest priority level from this Phase's review items
- Files with no review items are annotated `--` (no separate review needed)
- Sort by priority: P0 first, `--` last
### P0 — Must be human-reviewed
1. **Filename:line** — change description
Risk: why this needs attention
Suggestion: what to focus on during review
### P1 — Recommended attention
2. **Filename:line** — change description
Suggestion: checkpoints
### P2 — Quick scan
3. **Filename:line** — description
---
## Phase 2: ...P0/P1/P2 Assessment Criteria (consistent with /qq:brief):
- P0: Public interface changes, new cross-module dependencies, data format changes, state management/lifecycle changes, global static state isolation, anti-patterns (FindObjectOfType, etc.), resource cleanup/event unsubscription
- P1: Business logic branches, performance-sensitive paths (Update/FixedUpdate), O(N²) patterns, error handling/edge cases, new public methods or classes
- P2: Pure getters/setters/logging/comments, test code, config value tweaks
Key rule: Each review item appears only in the phase that introduced it — do not repeat it in later phases. If a later phase modifies code from an earlier phase, annotate it in that later phase as "modifies Phase X's ...".
6. Generate Document C: Review Guide
Generate REVIEW_GUIDE.md, indexing whichever documents exist in Docs/qq/<branch-name>/.
Format:
# Review Guide
> Branch: `<branch>`
> Generated: `<timestamp>`
## Document Index
| Document | Perspective | Purpose |
|----------|-------------|---------|
| `timeline-arch_<ts>.md` | Timeline × Architecture | Understand how the architecture evolved in development order |
| `timeline-review_<ts>.md` | Timeline × Review | Review code phase by phase, each with a file list |
| `arch-review_<ts>.md` | Final state × Architecture | Final architecture overview + module heatmap (if exists) |
| `pr-review_<ts>.md` | Final state × Review | Full final P0/P1/P2 list (if exists) |
If a document does not exist, annotate in the table "Not generated — run `/qq:brief` to generate".
## Reading Order
### Scenario A: First time looking at this branch
Use when: just picked it up, cross-team review, or returning after a long break
1. **timeline-arch** — follow the timeline to understand "how it got to this state", build a mental model
2. **arch-review** — view the final architecture overview and confirm the model is complete
3. **timeline-review** — review code phase by phase (use the quick reference to pick high-risk phases first)
4. **pr-review** — final scan from the final-state perspective to catch systemic risks across phases
### Scenario B: Already familiar with the branch, reviewing code directly
Use when: self-reviewing your own code, routine incremental review
1. **timeline-review** — quick reference → pick phases with most P0s → file list → review each item
2. **pr-review** — catch what was missed: the final-state perspective may reveal cross-phase combination risks
> Skip the two arch documents in Scenario B — you already know the architecture.
## Self-Review Workflow
Run /qq:timeline ↓ Open timeline-review, check the quick reference table ↓ Select phases in descending P0 count order (review highest-risk phases when most focused) ↓ For each Phase:
- Open the "Files to Review" table → open all P0 files
- Go through P0 items one by one → verify against the code
- Quickly scan P1 → only check items marked "Suggestion"
- Skip P2 (unless there's a question)
- Found an issue → fix it → commit ↓ All phases reviewed ↓ Make sure the project compiles, then run /qq:test to verify ↓ Optional: run /qq:brief to generate final-state docs for last-pass coverage ↓ Merge
## Time Budget
Summarize from the `timeline-review` quick reference table:
| Phase | Estimated Time |
|-------|---------------|
| (copied from quick reference) | |
| **Total** | **X min** |7. Output
Write three files to Docs/qq/<branch-name>/:
timeline-arch_<timestamp>.mdtimeline-review_<timestamp>.mdREVIEW_GUIDE.md(no timestamp, overwritten each time)
Branch name rule: use the current branch name, replacing / with _.
Timestamp format: YYYY-MM-DD-HHmm.
Notes
- Both timeline documents use identical Phase numbers and names
- A Phase with no points worth reviewing gets "No additional review needed for this phase" in the review document
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

