The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by processing untrusted external content as part of its core logic.
Ingestion points: Reads issue reports ({slug}-report.md) and arbitrary source code files discovered via search.
Boundary markers: Absent; there are no instructions to the agent to treat the content of these files as data only or to ignore embedded instructions.
Capability inventory: Performs code searching via Grep and Glob, and executes a local helper script search-yaml.py. The skill explicitly restricts code modification during this phase.
Sanitization: Absent; no validation or filtering of the ingested content is defined.
[COMMAND_EXECUTION]: Instructs the agent to use shell-based search utilities such as Grep and Glob, and to execute a local project script named search-yaml.py to analyze YAML-based metadata.

cs-issue-analyze