The Agent Skills Directory

[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by ingesting data from project-level files and interactive user input to generate test cases and drive remediation loops.
Ingestion points: Data is sourced from verification.json, validation.json, and .summaries/TASK-*.md in Step 4, as well as descriptive user feedback in Step 7.
Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions for the ingested content.
Capability inventory: The skill possesses significant capabilities, including file modification (Write/Edit), shell command execution (Bash), and the ability to spawn sub-agents (Agent tool).
Sanitization: There is no evidence of content validation or escaping before interpolation into agent prompts or execution plans.
[COMMAND_EXECUTION]: The skill uses the Bash tool for project file discovery in Step 2 and to execute build and sanity checks during the smoke testing phase in Step 3.
[COMMAND_EXECUTION]: In Step 12, the skill implements a 'gap-fix loop' that plans and executes automated fixes, which involves dynamic execution of commands based on LLM-generated plans derived from project state.

quality-test