The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted code and documentation, creating a vulnerability surface for indirect prompt injection.
Ingestion points: Files processed during Analyze Mode and Verify Mode, including user-specified paths and the production/qa/bugs/ directory.
Boundary markers: None. The skill does not use delimiters to isolate external data from its own instructions.
Capability inventory: Includes Read, Write, Glob, and Grep tools.
Sanitization: No validation or sanitization is performed on the content of the files read.
[COMMAND_EXECUTION]: The instructions for verifying bugs include an explicit command to execute shell scripts.
Evidence: "Run the related test — if the bug's system has a test file in tests/, run it via Bash and report pass/fail."
Configuration Conflict: This instruction attempts to use a shell execution tool that is not authorized in the skill's allowed-tools YAML frontmatter.
[DYNAMIC_EXECUTION]: The skill is designed to execute scripts found within the project structure dynamically.
Evidence: Automatically identifies and runs test files from the tests/ directory to confirm bug resolutions.

bug-report