The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and acts upon instructions (tasks and validation commands) found in external markdown files within the target repository.
Ingestion points: Reads fast-track-ai/tracks/NNN-slug/intent.md, spec.md, and tasks.md from the target repository.
Boundary markers: Requires status: approved frontmatter in spec.md and tasks.md before execution, but this is a semantic requirement rather than a security boundary for command execution.
Capability inventory: The skill has the capability to execute shell commands (Validate commands), write files within the target repository, and launch subagents.
Sanitization: No sanitization or sandboxing is mentioned for the validation commands extracted from the target repository's tasks.md file.
[COMMAND_EXECUTION]: The skill is designed to execute arbitrary commands defined in the project's task list.
Evidence: "Run each open task's Validate command and the checks the target project actually provides (such as existing test suites or typecheckers)" in Step 5.
[COMMAND_EXECUTION]: The skill executes specific management and gating scripts located in sibling directories relative to the skill's installation.
Evidence: ../fast-track-ai-spec/scripts/fast-track-ai-gate.sh in Step 7.
Evidence: ../fast-track-ai-init/scripts/fast-track-ai-index.sh in Step 8.
Context: These scripts are identified as "owner copy" (vendor-provided) and are part of the fast-track-ai tool suite managed by the author.

fast-track-ai-implement