The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes command-line tools to retrieve project history and status.
It executes git log to view recent commit history and divergence from origin.
It executes gh pr list to fetch pull request status from GitHub.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from repository files and external metadata, which creates a surface for indirect prompt injection.
Ingestion points: The agent is instructed to read contents from DESIGN.md, ROADMAP*, README, TODO*, and files within the .context/ directory, as well as git logs and pull request details.
Boundary markers: The skill does not define explicit delimiters or instructions to ignore embedded commands within the files it reads.
Capability inventory: The skill possesses the ability to execute shell commands (git, gh) and perform file read operations across the workspace.
Sanitization: There is no evidence of sanitization or filtering applied to the content synthesized from these external sources.
[SAFE]: Interaction with GitHub via the gh tool targets a well-known service for project metadata and does not involve exfiltration of sensitive local credentials.

recap