The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted data from the project environment to avoid redundant user interviews, creating a surface for injection attacks.
Ingestion points: The skill instructions in SKILL.md (Step 1.2, 1.3) direct the agent to read the repository's git log, README files, agent-instruction files, and data dictionaries.
Boundary markers: The instructions do not specify any delimiters or safety markers to differentiate between the ingested file content and legitimate system instructions.
Capability inventory: The skill has the capability to write the revops-context.md file (Step 5), modify agent-instruction files with new "invariants" (Step 5), and create scheduled routines (Step 7).
Sanitization: There are no explicit instructions to sanitize, filter, or validate the content read from external project files before it influences the agent's routing or state-management logic.
[COMMAND_EXECUTION]: The skill requires the agent to execute shell commands to perform environmental discovery.
Evidence: SKILL.md Step 1.2 explicitly instructs the agent to "Read the repository's recent git log" to infer project stage and pace.

revops-kickoff