github-elements-tracking
This skill should be used when the user asks to "track work across sessions", "create an epic", "manage issue waves", "post a checkpoint", "claim an issue", "recover from compaction", "coordinate multiple agents", "update memory bank", "store large documents", or mentions GitHub Issues as persistent memory, multi-session work, context survival, agent collaboration, SERENA MCP memory, or project-level context. Provides complete protocols for using GitHub Issues as permanent memory that survives context exhaustion, with integrated SERENA MCP memory bank for project-level context and large document storage.
What this skill does
## IRON LAW: User Specifications Are Sacred
**THIS LAW IS ABSOLUTE AND ADMITS NO EXCEPTIONS.**
1. **Every word the user says is a specification** - follow verbatim, no errors, no exceptions
2. **Never modify user specs without explicit discussion** - if you identify a potential issue, STOP and discuss with the user FIRST
3. **Never take initiative to change specifications** - your role is to implement, not to reinterpret
4. **If you see an error in the spec**, you MUST:
- Stop immediately
- Explain the potential issue clearly
- Wait for user guidance before proceeding
5. **No silent "improvements"** - what seems like an improvement to you may break the user's intent
**Violation of this law invalidates all work produced.**
## Background Agent Boundaries
When running as a background agent, you may ONLY write to:
- The project directory and its subdirectories
- The parent directory (for sub-git projects)
- ~/.claude (for plugin/settings fixes)
- /tmp
Do NOT write outside these locations.
---
## GHE_REPORTS Rule (MANDATORY)
**ALL agent reports MUST be posted to BOTH locations:**
1. **GitHub Issue Thread** - Full report text (NOT just a link!)
2. **GHE_REPORTS/** - Same full report text (FLAT structure, no subfolders!)
**Report naming:** `<TIMESTAMP>_<title or description>_(<AGENT>).md`
**Timestamp format:** `YYYYMMDDHHMMSSTimezone`
**Examples:**
- `20251206143000GMT+01_epic_15_wave_launched_(Athena).md`
- `20251206143022GMT+01_issue_42_dev_complete_(Hephaestus).md`
- `20251206150000GMT+01_issue_42_tests_passed_(Artemis).md`
- `20251206160000GMT+01_issue_42_review_complete_(Hera).md`
**ALL 11 agents write here:** Athena, Hephaestus, Artemis, Hera, Themis, Mnemosyne, Hermes, Ares, Chronos, Argos Panoptes, Cerberus
**REQUIREMENTS/** is SEPARATE - permanent design documents with legal validity, NEVER deleted.
**Deletion Policy:**
- GHE_REPORTS should be git-tracked - it constitutes the pulse of the GHE plugin
- DELETE ONLY when user EXPLICITLY orders deletion due to space constraints
- DO NOT delete during normal project cleanup or just because reports were archived to GitHub
---
## Project Settings
This skill respects settings in `.claude/ghe.local.md`. Run `/ghe:setup` to configure.
| Setting | Effect on This Skill |
|---------|---------------------|
| `enabled` | If false, skip all GitHub Elements operations |
| `enforcement_level` | strict/standard/lenient - affects rule strictness |
| `serena_sync` | If false, skip SERENA memory bank integration |
| `auto_worktree` | If true, auto-create git worktree on claim |
| `checkpoint_interval_minutes` | Reminder interval for checkpoints |
| `notification_level` | verbose/normal/quiet output |
**Defaults** (no settings file): enabled=true, enforcement=standard, serena_sync=true
---
## Related GHE Skills
For specific operations, GHE provides specialized skills:
| Skill | Purpose |
|-------|---------|
| **ghe-requirements** | Create, version, and link requirements files. Use when starting features. |
| **ghe-changelog** | Maintain version history with git-diff. Track code, requirements, and design changes. |
| **ghe-status** | Check thread status and context |
| **ghe-claim** | Claim a thread for work |
| **ghe-checkpoint** | Post progress checkpoints |
| **ghe-transition** | Request phase transitions |
| **ghe-report** | Generate status reports |
### Key Workflows
**Starting a New Feature:**
1. Use **ghe-requirements** to create REQ file
2. Link requirements to DEV issue
3. Use **ghe-claim** to claim the thread
4. Follow TDD workflow from DEV manager
**Tracking Changes:**
1. Use **ghe-changelog** after significant commits
2. Updates automatically track requirements and code changes
**Requirements-First Development:**
All DEV threads MUST link to a requirements file. The DEV manager (Hephaestus) enforces:
- Requirements breakdown into atomic changes
- TDD cycle for each atomic change (RED-GREEN-REFACTOR)
- Test coverage verification before transition
---
# GitHub Elements Tracking
GitHub Issues as **permanent memory** for AI agents. Survives compaction, enables collaboration, provides complete traceability.
**Integrated with SERENA MCP memory bank** for project-level context and large document storage beyond GitHub's limits.
---
## FUNDAMENTAL PRINCIPLES (READ FIRST)
These principles are non-negotiable. Violating them corrupts the entire workflow.
### 1. THE SACRED ORDER: One Branch, One Phase at a Time
Each branch goes through phases in strict circular order:
```
DEV ───► TEST ───► REVIEW ───► DEV ───► ... (until REVIEW passes)
│ │
│ └─► PASS? → merge to main
│
Bug fixes ONLY (no structural changes, no new tests)
```
**Critical Rules:**
- **ONE phase thread open at a time** per branch
- Opening TEST = closing DEV
- Opening REVIEW = closing TEST
- REVIEW fail → back to DEV (NEVER to TEST)
- REVIEW pass → merge to main
### 2. NEW BUG REPORT = NEW BRANCH (ALWAYS)
When a bug report is posted as a **new GitHub issue**:
- ALWAYS create a NEW branch with NEW DEV → TEST → REVIEW cycle
- NEVER merge it into an existing review thread, even if related
- Mixing issues violates the Sacred Order
**Exception**: Comments posted directly IN an existing thread are handled within that thread.
### 3. RESPONSIBILITY BOUNDARIES
| Agent | Handles Bug Reports? | Handles Test Writing? | Renders Verdicts? |
|-------|---------------------|----------------------|-------------------|
| **DEV** | NO (creates new branches for validated bugs) | YES | NO |
| **TEST** | NO (only runs existing tests) | NO | NO |
| **REVIEW** | YES (triages ALL bug reports) | NO | YES |
**REVIEW handles ALL bug reports and quality evaluation. TEST only runs existing tests.**
### 4. THE 3-STRIKE RULE FOR BUG REPORTS
When a bug report cannot be reproduced:
1. **Strike 1**: Politely ask for more details
2. **Strike 2**: Politely ask again with specific questions
3. **Strike 3**: Politely close as "cannot-reproduce"
**ALWAYS be polite**, even when closing. Thank the reporter for their effort.
### 5. EVERY CONTRIBUTION IS VALUED
- Never dismiss a contribution as "nitpick" or "pedantic"
- Verify everything, assume nothing
- If valid, thank the contributor regardless of size
- Reply respectfully even when rejecting
---
## Terminology
Understanding the relationship between GitHub entities and workflow concepts:
| Term | Definition |
|------|------------|
| **Issue** | A GitHub Issue - the container/ticket that holds all work history |
| **Thread** | A phase-specific issue. A "DEV thread" is an issue labeled `phase:dev` |
| **Comment** | A reply within an issue - holds checkpoints, decisions, work logs |
| **Element** | A comment containing Knowledge, Action, and/or Judgement facets |
**Key insight**: Every thread IS an issue. The terms describe the same GitHub entity from different perspectives:
- "Issue" emphasizes the GitHub container (the thing with a number like #201)
- "Thread" emphasizes the phase role (DEV, TEST, or REVIEW work)
**Example**: Issue #201 labeled `phase:dev` is the "DEV thread" for JWT authentication. When DEV completes, issue #201 closes and issue #202 (labeled `phase:test`) opens as the "TEST thread" for the same feature.
**One thread per phase**: Each phase (DEV, TEST, REVIEW) gets its own issue. When phase transitions occur, the old issue closes and a new issue opens with the new phase label.
### Thread Naming Convention
All thread titles MUST follow this standard format:
| Thread Type | Title Format | Example |
|-------------|--------------|---------|
| **DEV** | `[DEV] #ISSUE - Short Description` | `[DEV] #201 - JWT Authentication` |
| **TEST** | `[TEST] #ISSUE - Short Description` | `[TEST] #201 - JWT Authentication` |
| **REVIEW** | `[REVIEW] #ISSUE - Short Description` | `[REVIEW] #201 - JWT Authentication` |
| **DEV (iteration)** | `[DEV] #ISSUE - Description (Iteration N)` | `[DEV] #201 - JWT Auth (Iteration 2)` |
**RuleRelated in AI Agents
skill-development
IncludedComprehensive meta-skill for creating, managing, validating, auditing, and distributing Claude Code skills and slash commands (unified in v2.1.3+). Provides skill templates, creation workflows, validation patterns, audit checklists, naming conventions, YAML frontmatter guidance, progressive disclosure examples, and best practices lookup. Use when creating new skills, validating existing skills, auditing skill quality, understanding skill architecture, needing skill templates, learning about YAML frontmatter requirements, progressive disclosure patterns, tool restrictions (allowed-tools), skill composition, skill naming conventions, troubleshooting skill activation issues, creating custom slash commands, configuring command frontmatter, using command arguments ($ARGUMENTS, $1, $2), bash execution in commands, file references in commands, command namespacing, plugin commands, MCP slash commands, Skill tool configuration, or deciding between skills vs slash commands. Delegates to docs-management skill for official documentation.
reprompter
IncludedTransform messy prompts into well-structured, effective prompts — single or multi-agent. Use when: "reprompt", "reprompt this", "clean up this prompt", "structure my prompt", rough text needing XML tags and best practices, "reprompter teams", "repromptception", "run with quality", "smart run", "smart agents", multi-agent tasks, audits, parallel work, anything going to agent teams. Don't use when: simple Q&A, pure chat, immediate execution-only tasks. See "Don't Use When" section for details. Outputs: Structured XML/Markdown prompt, quality score (before/after), optional team brief + per-agent sub-prompts, agent team output files. Success criteria: Single mode quality score ≥ 7/10; Repromptception per-agent prompt quality score 8+/10; all required sections present, actionable and specific.
adaptive-compaction
IncludedAdaptive add-on policy and recovery layer that decides WHEN to compact, prune, snapshot, or fork -- replacing fixed-percent auto-compaction across Claude Code, Codex, and MCP-capable hosts. Trigger on auto-compact timing or damage: "when should I compact", "is it safe to compact now or start a fresh session", "auto-compact fires too early/mid-task", "switching to an unrelated task but the window still has space", "context rot", "answers get worse the longer the session runs", "the agent forgot the plan or my decisions after it summarized", "add a layer on top that manages context without changing the agent", raising autoCompactWindow to give the policy room, or installing/tuning a cross-tool compaction policy or PreCompact hook -- even when "compaction" is never said but the problem is context-window pressure or post-summarization memory loss. Do NOT use to summarize a conversation, build RAG, write a summarization prompt (decides WHEN not HOW), or answer max-context-length trivia.
agent-skill-creator
IncludedCreate cross-platform agent skills from workflow descriptions. Activates when users ask to create an agent, automate a repetitive workflow, create a custom skill, or need advanced agent creation. Triggers on phrases like create agent for, automate workflow, create skill for, every day I have to, daily I need to, turn process into agent, need to automate, create a cross-platform skill, validate this skill, export this skill, migrate this skill. Supports single skills, multi-agent suites, transcript processing, template-based creation, interactive configuration, cross-platform export, and spec validation.
llm-wiki
IncludedUse when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.
skill-master
IncludedAgent Skills authoring, evaluation, and optimization. Create, edit, validate, benchmark, and improve skills following the agentskills.io specification. Use when designing SKILL.md files, structuring skill folders (references, scripts, assets), ingesting external documentation into skills, running trigger evals, benchmarking skill quality, optimizing descriptions, or performing blind A/B comparisons. Keywords: agentskills.io, SKILL.md, skill authoring, eval, benchmark, trigger optimization.