The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/verification_context.py executes local git commands (status, diff, rev-parse) to collect metadata. The implementation uses subprocess.run with argument lists and shell=False, which is a secure method that prevents shell injection.
[PROMPT_INJECTION]: The skill contains a surface for indirect prompt injection because it processes untrusted data from the git repository and interpolates it into the agent's instructions.
Ingestion points: scripts/verification_context.py gathers file names, branch names, and diff statistics from the local repository environment.
Boundary markers: Absent; the references/prompt-template.md uses a simple <verification_context> placeholder without delimiters or explicit instructions to treat the content as data only.
Capability inventory: The skill allows the agent to read repository metadata and perform a verification analysis.
Sanitization: No sanitization or validation is performed on the data retrieved from the git repository before it is passed to the language model.

verification-gate