agentleFS
Ron-daliRon-dali/vibe-coding-rules/changelog/SKILL.md

changelog

Skill4 starsChanged 4 months agoApache-2.0View on GitHub

代码修改→自检通过→测试通过后的最后一步:将改动写入项目变更日志。 Final step after code changes complete, self-check passes, and tests pass: write changes to project changelog. 支持自增长检测(发现重复Bug→自动升级规则)、反膨胀清理、一键分享导出。 Strictly follow this template, writing to {{pipeline.paths.changelogFormat}}: Before writing the changelog, execute self-growth detection. See references/growth-rules.md for detailed rule definitions. Execute _detectGrowthOpportunity(): Why: Without cleanup, rules accumulate indefinitely. A 6-month project can have 100+ rules, bloating every self-check. This step keeps the rule set lean. Execute _runAntiBloatCheck() per references/growth-rules.md — after growth detection, before changelog write. For each [SOFT] rule in self-check-full.md,…

  • Installs packages

160 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use playtest-report, soak-test and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • CoralGameSkill

    playtest-report

    Generates a structured playtest report template or analyzes existing playtest notes into a structured format. Use this to standardize playtest feedback collection and analysis.

    Also in 12 projects that use this

  • lilangMaxSkill

    soak-test

    Generate a soak test protocol for extended play sessions. Defines what to observe, measure, and log during long play sessions to surface slow leaks, fatigue effects, and edge cases that only appear after sustained play. Primarily used in Polish and Release phases.

    Also in 9 projects that use this

  • lilangMaxSkill

    team-live-ops

    Orchestrate the live-ops team for post-launch content planning: coordinates live-ops-designer, economy-designer, analytics-engineer, community-manager, writer, and narrative-director to design and plan a season, event, or live content update.

    Also in 9 projects that use this

  • CoralGameSkill

    team-polish

    Orchestrate the polish team: coordinates performance-analyst, technical-artist, sound-designer, and qa-tester to optimize, polish, and harden a feature or area for release quality.

    Also in 14 projects that use this

  • +5 more from lilangMax
  • +3 more from CoralGame

The file

Show its rules and text

We split this skill into the rules it gives an agent

20 rules. 0 are common, 9 are distinctive, and 11 are very specific to this project.

  • 0

    Common

    Most of the 149 edited versions we read keep these.

  • 9

    Distinctive

    This author's own choices. Worth a look.

  • 11

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Not exists → create new observation entry O-00X, confidence = 1

    line 62
  2. Distinctive:

    On completion: Write "Growth Detection" entry in changelog summary

    line 70
  3. Very specific:

    If not triggered for N sessions (config: softStalenessSessions, default 10) → demote to observation pool, confidence reset to 1

    line 81
  4. Very specific:

    If total > maxTotalRules (default 30): demote lowest-confidence [SOFT] rules until under cap

    line 87
  5. Very specific:

    If [HARD] > maxHardRules (default 15): demote lowest-confidence [HARD] → [SOFT] until under cap

    line 88
  6. Very specific:

    If observations > maxObservations (default 50): remove stale entries with confidence=1 and no re-trigger in 30 days

    line 92
  7. Distinctive:

    If still over: remove lowest-confidence entries

    line 93
  8. Very specific:

    If retired > retiredArchiveThreshold (default 20): distill oldest 10 into a summary row

    line 97
  9. Distinctive:

    If no actions: "Anti-bloat check: rule set is within healthy limits."

    line 114
  10. Very specific:

    Use current date, format YYYY-MM-DD.

    line 117
  11. Very specific:

    Trigger: When this session has Tier 2 [HARD] rules generated or promoted AND shareContact.phone is not empty.

    line 132
  12. Very specific:

    Before outputting: Check pipeline.json → selfGrowth.shareContact.phone.

    line 134
  13. Distinctive:

    Only list rules promoted or triggered as [HARD] this session

    line 156
  14. Distinctive:

    If no [HARD] rule changes this session → skip regardless

    line 160
  15. Very specific:

    Read project overview doc (check root OVERVIEW.md first, then README.md)

    line 166
  16. Very specific:

    Search for version number references (patterns like v1.0.0, version 1.0)

    line 167
  17. Distinctive:

    If hardcoded version numbers exist → replace with new version

    line 168
  18. Very specific:

    If overview doc references .ai-pipeline/VERSION → skip (already using best practice)

    line 169
  19. Distinctive:

    If project has no overview doc → skip

    line 170
  20. Distinctive:

    Do not skip: Must write after self-check + tests pass.

    line 177

“Common” is measured: kept by at least 6 in 10 of the 149 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Changelog / 变更日志

Overview / 概述

代码修改→自检通过→测试通过后的最后一步:将改动写入项目变更日志。

Final step after code changes complete, self-check passes, and tests pass: write changes to project changelog. 支持自增长检测(发现重复Bug→自动升级规则)、反膨胀清理、一键分享导出。

Write Format

Strictly follow this template, writing to {{pipeline.paths.changelogFormat}}:

# Changelog - YYYY-MM-DD

## Change Summary
One sentence describing what was done (future-facing, explain "why")

---

## Changed Files

| File Path | Change Type | Brief Description |
|:--|:--|:--|
| `path/to/file.js` | Added/Modified/Deleted | Description |

---

## New Dependencies

None (or list package name + version)

---

## Self-Check Results Digest

| Check Item | Result |
|:--|:--|
| Architecture registration consistency | N/A / Passed |
| Schema field completeness | N/A / Passed |
| Truncation point scan | N/A / Passed |
| Numeric `\|\|` falsy check | N/A / Checked |
| `<select>` placeholder | N/A / Checked |
| POST upsert hardcoded values | N/A / Checked |
| Async race conditions | N/A / Checked |
| Lint | All passed |

Workflow

Step 0: Growth Detection (V2.1+)

Before writing the changelog, execute self-growth detection. See references/growth-rules.md for detailed rule definitions.

Execute _detectGrowthOpportunity():

  1. Scan this session: Review all replace_in_file / write_to_file operations
  2. Identify fix iterations: Same file modified multiple times + bug fix signals ("fix", "bug", "error", "still broken")
  3. Extract bug pattern: Distill general pattern from user feedback and fix content
  4. Match with observation pool (references/observations.md):
    • Exists → confidence +1, record reoccurrence
    • Not exists → create new observation entry O-00X, confidence = 1
  5. Check promotion thresholds (refer to pipeline.json selfGrowth.confidenceThresholds):
    • Confidence ≥ 3 → promote to [SOFT] rule, append to self-check-full.md auto-grown rules zone
    • Confidence ≥ 5 → promote to [HARD] rule
    • Record promotion event in growth-log.md
  6. Check retirement conditions:
    • User rejected the rule suggestion during this session → confidence -2
    • Confidence drops to 0 → move to retired.md
  7. On completion: Write "Growth Detection" entry in changelog summary

Step 0.5: Anti-Bloat Check (V2.2+)

Why: Without cleanup, rules accumulate indefinitely. A 6-month project can have 100+ rules, bloating every self-check. This step keeps the rule set lean.

Execute _runAntiBloatCheck() per references/growth-rules.md — after growth detection, before changelog write.

1. Staleness Sweep (SOFT rules)

For each [SOFT] rule in self-check-full.md, check growth-log.md for last_triggered:

  • If not triggered for N sessions (config: softStalenessSessions, default 10) → demote to observation pool, confidence reset to 1
  • Log event in growth-log.md

2. Hard Cap Enforcement

Count [SOFT] + [HARD] rules:

  • If total > maxTotalRules (default 30): demote lowest-confidence [SOFT] rules until under cap
  • If [HARD] > maxHardRules (default 15): demote lowest-confidence [HARD] → [SOFT] until under cap

3. Observation Pool Pruning

  • If observations > maxObservations (default 50): remove stale entries with confidence=1 and no re-trigger in 30 days
  • If still over: remove lowest-confidence entries

4. Retired Archives Compaction

  • If retired > retiredArchiveThreshold (default 20): distill oldest 10 into a summary row

5. Write Anti-Bloat Report

Append to changelog:

## Anti-Bloat Report (V2.2)

| Action | Details |
|:--|:--|
| Staleness sweep | N SOFT rules demoted |
| Hard cap enforcement | N rules demoted |
| Observation pruning | N old observations removed |
| Retired compaction | N retired entries distilled |

If no actions: "Anti-bloat check: rule set is within healthy limits."

Step 1: Determine Date

Use current date, format YYYY-MM-DD.

Step 2: Collect Info

  • Change summary: User requirement + distilled change description
  • Changed file list: git status or actual modification list
  • New dependencies: npm install package names, "None" if empty
  • Self-check results: Fill item by item from self-check output

Step 3: Read/Write File

  • File exists → replace_in_file append new change block
  • File doesn't exist → write_to_file create new file
  • Same day multiple changes: append to same file, separated by ---

Step 4: One-Click Share Export (V2.1+)

Trigger: When this session has Tier 2 [HARD] rules generated or promoted AND shareContact.phone is not empty.

Before outputting: Check pipeline.json → selfGrowth.shareContact.phone. If empty → skip this entire block.

Append share text block at end of changelog:

---
## 💡 Important Rules Auto-Discovered by AI

During this development session, the self-growing system discovered the following Tier 2 [HARD] rules:

> **[HARD] Rule Name**: Rule description
> Trigger: XXX
> Correct Pattern: XXX
> Confidence: N (consistently triggered N times)

---
These rules come from the **Vibe Coding Rules V2.2** self-growing system.
📱 Share: {{pipeline.selfGrowth.shareContact.phone}}
({{pipeline.selfGrowth.shareContact.projectName}})

Format requirements:

  • Only list rules promoted or triggered as [HARD] this session
  • One line per rule with brief description + confidence
  • Phone and project name read from pipeline.json selfGrowth.shareContact
  • If phone is empty → skip this block entirely (opt-in privacy design)
  • If no [HARD] rule changes this session → skip regardless

Step 5: Sync Overview Doc After Version Bump (V2.2+)

When this change triggers a VERSION bump, synchronize the project overview document (README.md / OVERVIEW.md):

  1. Read project overview doc (check root OVERVIEW.md first, then README.md)
  2. Search for version number references (patterns like v1.0.0, version 1.0)
  3. If hardcoded version numbers exist → replace with new version
  4. If overview doc references .ai-pipeline/VERSION → skip (already using best practice)
  5. If project has no overview doc → skip

Purpose: Prevent churn — eliminate AI-generated zombie version numbers from inconsistent multi-file maintenance.

Key Constraints

  • Write "why changed" not "what changed": Git tracks file changes, changelog tracks business reasons
  • Do not skip: Must write after self-check + tests pass. No "I'll do it later."

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.