The Agent Skills Directory

[COMMAND_EXECUTION]: The agent is instructed to execute a local data collection script (scripts/collect/week.ts) using npx tsx and to perform workspace diagnostics using the cargo-ai CLI tool. These commands are used to gather evidence regarding git activity, deployed infrastructure, and run metrics.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external repository sources which could contain malicious instructions designed to influence the agent's recap or recommendations.
Ingestion points: Data enters the agent's context through git log and gh pr list outputs, as well as content from files in the initiatives/ and cadence/ directories.
Boundary markers: The instructions do not define clear delimiters or include warnings for the agent to ignore instructions embedded within the processed git or pull request data.
Capability inventory: The agent has capabilities to execute shell commands (cargo-ai), write files to the repository, and perform network-enabled operations (creating branches and opening pull requests) via the GitHub connector.
Sanitization: The skill does not implement sanitization or filtering for the data collected from git history or pull request metadata before presenting it to the language model.

weekly-planning