The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests existing C# source code files to identify test scenarios and generate code. Maliciously crafted content within these files could influence the agent's logic or the instructions sent to the sub-agent.
Ingestion points: The skill uses the Read tool in Phase 1 to process the file path provided in the arguments.
Boundary markers: The prompt instructions do not specify any delimiters or safety warnings to ignore instructions embedded within the source files.
Capability inventory: The skill has the ability to Write and Edit files, execute shell commands via Bash, and delegate tasks to sub-agents via Task.
Sanitization: There is no evidence of sanitization or validation performed on the ingested code before it is passed into the analysis phase.
[COMMAND_EXECUTION]: The skill executes a shell command using Node.js to run a script located at a relative path in a sibling directory.
Evidence: node skills/csharp-tdd-develop/scripts/test-detector.js --detect
Observation: This creates a execution dependency on the local file system and assumes the presence and security of the csharp-tdd-develop skill scripts.

csharp-test-develop