The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to run various shell commands, including local scripts and standard development tools, which take user-provided file paths as arguments.
Evidence: SKILL.md directs the agent to execute python3 scripts/resolve_style_guides.py <changed-path>... and python3 scripts/validate_trigger_matrix_sync.py.
Evidence: references/quality-gate-command-matrix.md contains a matrix of commands for tools like dotnet, ruff, cargo, sqlfluff, and terraform to be run by the agent.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because its primary function is to ingest and process untrusted code and configuration files.
Ingestion points: Files matching patterns in references/trigger-matrix.md (e.g., .cs, .sh, .py, .js, .tf, .sql) are loaded into the agent's context.
Boundary markers: Absent. There are no instructions to use delimiters or ignore instructions embedded within the analyzed files.
Capability inventory: The skill has shell execution capabilities (dotnet, python3, etc.) which could be exploited if an attacker injects malicious instructions into the code being reviewed.
Sanitization: Absent. The skill does not mention any validation or escaping of the content of the files before they are processed by the LLM.

csharp-style-guide