The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute various subcommands of the cawplan and qa-insights CLI utilities. These commands are used to fetch mappings, list sections, and link artifacts between CawPlan and TestRail. This execution is restricted to the specific, task-oriented subcommands defined in the allowed commands table.- [INDIRECT_PROMPT_INJECTION]: The skill processes external data such as TestRail project URLs, suite IDs, and case metadata. While this ingestion represents a surface for indirect prompt injection, the risk is mitigated by the skill's requirement for human-in-the-loop confirmation gates (the --confirm flag) before any write operations are performed. The agent is also instructed to parse specific IDs rather than broad natural language descriptions when constructing commands.- [DYNAMIC_EXECUTION]: The skill generates local JSON body files to pass complex data structures to the CLI tools (e.g., using --body-file). This is a standard and secure method for providing structured input to command-line interfaces, avoiding the risks associated with raw string interpolation of potentially untrusted metadata into shell commands.

cawplan-testrail-link