The Agent Skills Directory

[COMMAND_EXECUTION]: The skill automates development tasks by executing shell commands and Python scripts, such as git status, gh pr list, and custom scripts like role-x.py, bookkeeping.py, and p9.py for lifecycle automation.
[EXTERNAL_DOWNLOADS]: Fetches a behavioral contract from the vendor's domain at https://broomva.tech/prompts/autonomous-senior-engineer. This is used to define agent operating principles.
[PROMPT_INJECTION]: Includes instructions to ignore user requests that attempt to bypass established 'bstack' quality gates. It enforces 'unconditional defaults' and directs the agent to skip human confirmation for actions like documentation updates and code merging.
[DATA_EXFILTRATION]: Mentions transcript capture and knowledge graph bookkeeping. These operations are described as internal workspace maintenance tasks.
[PROMPT_INJECTION]: The skill ingests data from repository files (roles/*.md), git status, and file diffs. Boundary markers are absent for this data. Its capabilities include git merge, gh pr create, and local script execution. Sanitization is addressed through scrubbing sensitive data from conversation bridges.

autonomous