Update to AI DevKit

{
  "name": "memory",
  "description": "AI DevKit · Use the memory CLI as a durable knowledge layer. Search before non-trivial work, store verified reusable knowledge, update stale entries, and avoid saving transcripts, secrets, or one-off task progress.",
  "included_files": [
    {
      "relative_path": "agents/openai.yaml",
      "size_in_bytes": 232
    }
  ],
  "skill_md_contents": "---\nname: memory\ndescription: AI DevKit · Use the memory CLI as a durable knowledge layer. Search before non-trivial work, store verified reusable knowledge, update stale entries, and avoid saving transcripts, secrets, or one-off task progress.\n---\n\n# AI DevKit Memory CLI\n\nUse `npx ai-devkit@latest memory ...` as the durable knowledge layer.\n\n## Workflow\n\n1. For implementation, debugging, review, planning, or documentation tasks, search before deep work unless the task is trivial:\n   ```bash\n   npx ai-devkit@latest memory search --query \"<task, subsystem, error, or convention>\" --limit 5\n   ```\n   For broad or risky tasks, search multiple angles: subsystem, error text, framework, command, and task intent.\n\n2. Use results as context:\n   - Trust repo files, tests, fresh command output, and explicit user instructions over memory.\n   - If memory conflicts with verified evidence, use the evidence and update the stale memory.\n   - Mention memory only when it changes the plan or avoids asking the user again.\n\n3. Search before storing:\n   ```bash\n   npx ai-devkit@latest memory search --query \"<knowledge to store>\" --table\n   ```\n\n4. Store or update only after the quality gate passes.\n\n## Quality Gate\n\nBefore storing, all must be true:\n\n- Future sessions are likely to reuse it.\n- It is verified by code, docs, tests, command output, or explicit user instruction.\n- It is not merely a restatement of obvious nearby files unless it prevents repeated agent mistakes.\n- It is scoped narrowly enough.\n- Existing memory does not already cover it.\n- It contains no secrets, credentials, private customer data, personal data, raw logs, or temporary paths.\n\nStore:\n- Project conventions, user preferences, durable decisions.\n- Reusable fixes, testing patterns, commands, setup gotchas.\n- Non-obvious constraints, architecture rules, failure patterns.\n\nDo not store:\n- Task progress, transcripts, speculation, generic programming facts.\n- Raw errors without diagnosis.\n- Anything the user did not intend to persist.\n\n## Commands\n\n### Search\n\n```bash\nnpx ai-devkit@latest memory search \\\n  --query \"<query>\" \\\n  --tags \"<tags>\" \\\n  --scope \"<scope>\" \\\n  --limit 5\n```\n\nUse `--table` to get IDs for updates:\n\n```bash\nnpx ai-devkit@latest memory search --query \"<query>\" --table\n```\n\nOptions: `--query/-q` required; `--tags`; `--scope/-s`; `--limit/-l` from 1-20; `--table`.\n\n### Store\n\n```bash\nnpx ai-devkit@latest memory store \\\n  --title \"<actionable title, 10-100 chars>\" \\\n  --content \"<context, guidance, evidence, exceptions>\" \\\n  --tags \"<lowercase,tags>\" \\\n  --scope \"<global|project:name|repo:org/repo>\"\n```\n\nUse this content shape when helpful:\n\n```text\nContext: Where this applies.\nGuidance: What to do.\nEvidence: File, command, test, or user instruction.\nExceptions: When not to apply it.\n```\n\n### Update\n\nFind the ID with `search --table`, then update only changed fields:\n\n```bash\nnpx ai-devkit@latest memory update \\\n  --id \"<memory-id>\" \\\n  --title \"<updated title>\" \\\n  --content \"<updated content>\" \\\n  --tags \"<replacement,tags>\" \\\n  --scope \"<updated scope>\"\n```\n\n`--tags` replaces all existing tags.\n\n## Scoping\n\nUse the narrowest useful scope:\n\n- `repo:<org/repo>` for one repository.\n- `project:<name>` for one app, product, or workspace.\n- `global` only for knowledge that applies across unrelated projects.\n\nIf unsure, use a narrower scope.\n\n## Troubleshooting\n\n- CLI missing: run `npx ai-devkit@latest --version`.\n- Duplicate title: search, then update the existing item if it is the same knowledge.\n- Empty results: broaden terms, remove filters, or search symptoms and subsystem names separately.\n- Validation error: check title/content lengths, query length, and `--limit` range.\n- DB path: default is `~/.ai-devkit/memory.db`; project config can override it automatically.\n"
}