The Agent Skills Directory

[SAFE]: The skill performs expected developer actions such as reading git logs and diffs to understand code changes. These operations are essential for its stated purpose of creating UAT documentation.
[SAFE]: The skill produces text-based outputs (TSV, SQL, Markdown) and writes them to a specific local directory (uat/). It does not attempt to execute the generated SQL or external commands automatically, nor does it perform any network exfiltration.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an inherent attack surface for indirect prompt injection because it is designed to ingest and process data from external sources including user-provided specifications and project documentation (e.g., README, CLAUDE.md). While this is a standard design for such a tool, the risk is mitigated as the primary output is static documentation.
Ingestion points: Feature intent, pseudo-specifications, existing UAT artifacts, and repository files (README, docs, etc.) via the local file system and git commands.
Boundary markers: None identified; the skill does not explicitly prompt the agent to ignore instructions embedded within the ingested source material.
Capability inventory: Execution of git log and git diff; file system write access to the uat/ directory.
Sanitization: Not explicitly implemented in the skill instructions.

uat-scenarios