The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes content from codebase files during its discovery phase (Step 2) using tools like grep, glob, and grepika. This creates an attack surface where malicious instructions embedded in the files being analyzed could influence the agent's behavior.
Ingestion points: Codebase files and search snippets accessed via search tools in SKILL.md.
Boundary markers: The skill does not define specific boundary markers for the data it reads, although the final output is delimited in a fixed block format.
Capability inventory: The skill utilizes file system discovery (search) tools but does not explicitly perform file writes, network operations, or direct code execution.
Sanitization: No explicit sanitization or filtering of the processed file content is described.
[COMMAND_EXECUTION]: The skill instructs the agent to utilize various search tools, including grep, glob, and MCP-based tools like grepika and graphify, to explore the repository structure. These commands are used for discovery and are consistent with the skill's primary function.
[PROMPT_INJECTION]: The instructions include directives to "Skip silently" for certain tasks and to "proceed straight into the task... no confirmation gate." While intended to improve efficiency by reducing noise and interaction latency, these instructions reduce human oversight of the agent's actions during the initial scoping and implementation phases.

auto-scope