The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses common shell utilities including ls, mkdir, grep, and wc via the Bash tool to analyze project structure and manage documentation files in the docs/adrs/ directory.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its analysis of untrusted codebase data. Ingestion points: The skill reads project structure, configuration files such as package.json and tsconfig, and existing ADR markdown files. Boundary markers: No specific markers or instructions exist to isolate project content from agent instructions, which could allow malicious comments in source code to influence the agent's behavior. Capability inventory: The agent has access to the Bash tool for local command execution and the Write tool for file system modification. Sanitization: Content extracted from analyzed files is used directly in reasoning and document generation without explicit validation or escaping.

blueprint-derive-adr