The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions (references/ARL-agent-instructions.md) direct the agent to perform git clone operations on six specific external GitHub repositories for research and analysis. These include https://github.com/bmad-code-org/BMAD-METHOD, https://github.com/steveyegge/gastown, https://github.com/glittercowboy/get-shit-done, https://github.com/bgauryy/octocode-mcp, https://github.com/mikeyobrien/ralph-orchestrator, and a vendor-owned resource at https://github.com/bitflight-devops/stateless-agent-methodology. The repositories are downloaded for analysis by specialized agent experts.
[INDIRECT_PROMPT_INJECTION]: The skill defines a research methodology involving the analysis of external codebases. This creates a potential attack surface for indirect prompt injection.
Ingestion points: The agent is instructed to clone and read the full source code of six external GitHub repositories (references/expert-repos.md).
Boundary markers: None are explicitly specified; the agent is tasked with a deep-dive analysis of the ingested content.
Capability inventory: The skill uses the agent to perform source-code surveys, expert cross-examination, and synthesis of research findings.
Sanitization: None; the agent processes the raw source code of the target frameworks.
[NO_CODE]: The skill contains only Markdown documentation and instructional materials. It does not include executable scripts, configuration files that could trigger package installation (like package.json), or platform-specific shell command injections.

arl