The Agent Skills Directory

[SAFE]: The skill is a planning assistant that performs architectural and implementation analysis within a repository context. It does not perform network operations or handle sensitive credentials.\n- [SAFE]: The skill contains explicit constraints to only perform analysis and never edit product files, establishing a clear safety boundary for its operations.\n- [SAFE]: The provided REFERENCE.md includes a dedicated 'Security/secrets' lens, directing the agent to proactively check for risks such as credential exposure, PII leaks, and new injection surfaces.\n- [INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it reads and processes external data from repository files (code, ADRs, and memory files). However, the risk is minimal as the skill is restricted to generating text-based plans and does not execute the code it analyzes.\n
Ingestion points: Repository source code, Architecture Decision Records (ADR) trees, and recon memory files.\n
Boundary markers: The instructions do not define specific delimiters or warnings for the agent when processing these files.\n
Capability inventory: The skill allows reading repository files and writing the resulting plan to a user-specified file.\n
Sanitization: The skill does not implement specific sanitization or filtering for the data it ingests from the codebase.

autoplan