robotaitai/project-bedrock/project-memory-writing
Claude Code and agent skill by robotaitai/project-bedrock
- community
Checking that the SKILL.md file is still in the repository.
Write high-quality, stable memory notes. Use when creating or updating notes in agent-knowledge/Memory/ and you want specific guidance on format, quality, and what to include.
Install
npx vskill@latest install robotaitai/project-bedrock/project-memory-writingSKILL.md
Project Memory Writing
Focused rules for writing durable memory notes. This is the craft side of memory management: how to write notes that stay useful across sessions and agents.
When to use this skill
- Creating a new branch note or subtopic note
- Updating Current State or Recent Changes sections
- Reviewing a note that has grown noisy or stale
- Any time the question is "how should I write this?"
For tree structure and reading strategy, see memory-management. For bootstrapping from scratch, see project-ontology-bootstrap.
Verification first
Before writing anything to Memory/, verify the fact against the source:
| Source type | Verification requirement | |
|---|---|---|
| Code file | Read the actual file, not a cached summary | |
| Test | Run it or check its current state in the test output | |
| Decision | Confirm it is still active (not reversed) | |
| Config | Read the current config file, not old docs | |
| Agent output | Treat as hypothesis until confirmed |
Do NOT write:
- Speculative plans
- In-progress task state (use Sessions/)
- Raw agent summaries without verification
- Facts that change faster than once per week
Required frontmatter
Every Memory/ note must open with YAML frontmatter:
note_type values: durable-memory-root (for MEMORY.md), branch-entry, branch-leaf, decision.
Do not add extra frontmatter fields that tools or other agents don't need.
Required sections for branch notes
Use this section order. Keep each section focused.
Quality rules
Be specific. "Uses Python 3.12" not "uses modern Python".
State facts, not observations. "Auth uses JWT with 24h expiry" not "I noticed auth seems to use JWT".
Stay bounded. Each note covers one domain. Do not add cross-domain facts to a branch note. If a fact belongs to two domains, put it in the more specific one and link from the other.
Keep Current State evergreen. Remove outdated bullets. Do not append without pruning. If you're adding a new fact, check if an old bullet should be removed.
Keep Recent Changes short. 6-10 entries. Oldest entries get pruned by compaction. The purpose is "what changed recently", not a full changelog.
Keep notes scannable. Use short bullets over prose paragraphs. Headers are navigation -- keep them consistent across notes.
Note sizing
| Size | Action | |
|---|---|---|
| Under ~100 lines | Fine. Keep as-is. | |
| 100-150 lines | Consider pruning Recent Changes. | |
| Over 150 lines | Split: create a folder + same-name entry note, move detail to a subtopic note. | |
| Over 200 lines | Split immediately. |
When splitting, the entry note stays as the summary and router. Subtopic notes hold the detail.
Linking
Use relative markdown links, not wiki-links:
Not:
Links from entry notes to subtopic notes:
Links from branch notes to decisions:
What goes where
| Content type | Location | |
|---|---|---|
| Stable, verified facts about the project | Memory/<branch>.md | |
| Raw imported content (web, docs, git log) | Evidence/raw/ or Evidence/imports/ | |
| Captured event history | Evidence/captures/ | |
| Generated views (indexes, exports) | Outputs/ | |
| Current session state | Sessions/ | |
| Architectural decisions with rationale | Memory/decisions/ |
Never put raw content in Memory/. Never put curated facts in Evidence/.
Memory/MEMORY.md rules
The root index must stay short (under 30 lines of content):
- One line per branch, with a brief inline summary
- No dense detail
- Updated only when branches are added, removed, or substantially renamed
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingassets/skills/project-memory-writing/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/robotaitai/project-bedrock/project-memory-writing)