# Codebase Second Brain

> Retrieve, verify, and refresh the persistent Open Knowledge Format v0.2 codebase memory for playwright-agentic-automation. Use when an agent needs architecture discovery, source ownership, page/API/fixture relationships, static AST graph queries, knowledge freshness checks, Mermaid/Obsidian navigation, or durable knowledge updates after repository changes.

- **Type:** Skill
- **Install:** `agentstack add skill-mallikarjun-roddannavar-playwright-agentic-automation-codebase-second-brain`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [Mallikarjun-Roddannavar](https://agentstack.voostack.com/s/mallikarjun-roddannavar)
- **Installs:** 0
- **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [Mallikarjun-Roddannavar](https://github.com/Mallikarjun-Roddannavar)
- **Source:** https://github.com/Mallikarjun-Roddannavar/playwright-agentic-automation/tree/main/.agents/skills/codebase-second-brain

## Install

```sh
agentstack add skill-mallikarjun-roddannavar-playwright-agentic-automation-codebase-second-brain
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Codebase Second Brain

Use `knowledge/` as the first discovery layer. It is a portable, offline-first OKF v0.2 bundle; `generated/` contains deterministic AST facts and `architecture/`, `decisions/`, and `runbooks/` hold durable human-readable knowledge.

## Retrieve Before Rediscovering

1. Read `knowledge/index.md`.
2. Run `npm run knowledge:check`. If it reports stale artifacts, treat the source code as authoritative and refresh before relying on generated facts.
3. Run `npm run knowledge:query -- `.
4. Open only the returned concept notes and their direct links. Read source only to verify an important claim, resolve a stale fact, or make the requested change.

Use `--relation` for exact relationship retrieval:

```bash
npm run knowledge:query -- --relation NAVIGATES_TO
npm run knowledge:query -- --relation USES_API_ROUTE
npm run knowledge:query -- --relation USES_FIXTURE
```

## Windows Shell Fallback

If `npm run knowledge:check` hits the known Windows `EPERM`/realpath issue, run the underlying commands directly:

```bash
node ./scripts/buildKnowledge.mjs --check
node ./scripts/validateKnowledge.mjs
```

For a direct query, omit npm's extra delimiter: `node ./scripts/queryKnowledge.mjs LoginPage`.

## Trust the Right Layer

- Treat TypeScript source, config, and tests as authoritative.
- Treat `knowledge/generated/code-graph.json` as machine-confirmed static analysis with source-hash provenance.
- Do not describe graph edges as runtime behavior. They are static imports, declarations, resolved inheritance, concrete `new` expressions, type returns, route references, fixtures, and package usage.
- Preserve unknown OKF frontmatter fields and tolerate broken Markdown links when consuming knowledge; do not silently invent missing facts.
- Load `references/static-graph-model.md` only when changing the extractor or interpreting an edge type.

## Refresh After a Relevant Change

When changing indexed TypeScript, JavaScript, JSON configuration, package metadata, or knowledge scripts:

1. Run `npm run knowledge:build`.
2. Run `npm run knowledge:validate`.
3. Review the generated concepts, `knowledge/generated/code-graph.json`, and Mermaid graphs for factual changes.
4. Keep the generated artifacts with the code change. Do not hand-edit files under `knowledge/generated/`.
5. Add human reasoning, decisions, exceptions, and runbooks outside `generated/`, using valid OKF frontmatter with a non-empty `type`.

## Obsidian

Open `knowledge/` as an Obsidian vault when a visual link graph or Mermaid rendering is useful. Use normal Markdown links so the bundle stays portable to any Markdown/OKF consumer. Do not add user-specific `.obsidian` settings to the repository unless explicitly requested.

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [Mallikarjun-Roddannavar](https://github.com/Mallikarjun-Roddannavar)
- **Source:** [Mallikarjun-Roddannavar/playwright-agentic-automation](https://github.com/Mallikarjun-Roddannavar/playwright-agentic-automation)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-mallikarjun-roddannavar-playwright-agentic-automation-codebase-second-brain
- Seller: https://agentstack.voostack.com/s/mallikarjun-roddannavar
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.