The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill uses dynamic context injection to execute a shell script (yaml-helper.sh) at load time. This is used to resolve configuration settings such as review_mode and team.size from the project environment.
[PROMPT_INJECTION]: The instructions explicitly direct the agent to bypass the 'Collaboration Protocol'—a standard requirement for user confirmation before writing files. The skill defines a 'bounded exception' allowing sub-agents to write new artifacts to production/, docs/, or tests/ without individual prompts, provided the phase itself was approved.
[INDIRECT_PROMPT_INJECTION]: The skill takes user input describing a feature or area and propagates it to sub-agents, creating an attack surface for indirect injection.
Ingestion points: Skill argument [feature or area to polish] in SKILL.md.
Boundary markers: Absent; user input is interpolated into briefs without delimiters.
Capability inventory: The pipeline uses Agent, Bash, Write, and Edit tools.
Sanitization: No input validation or sanitization is performed on the user-provided string.
[COMMAND_EXECUTION]: The skill performs command execution via the Bash tool and orchestrates multiple sub-agents to carry out profiling, testing, and implementation tasks. Note that the Bash tool usage is restricted in the frontmatter to a specific script and command.

team-polish