The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process content from arbitrary documentation URLs provided by the user via the Firecrawl agent. This represents an indirect prompt injection attack surface, as malicious instructions or adversarial content embedded within the external documentation could influence the agent's behavior during the skill generation process.
Ingestion points: The documentation URLs are passed to the mcp__firecrawl__firecrawl_agent tool as specified in the SKILL.md instructions.
Boundary markers: The skill utilizes a structured prompt template (references/skill-prompt-template.md) and a JSON schema to guide the extraction, but it does not include explicit safety markers to ignore embedded instructions in the source documentation.
Capability inventory: The agent has the capability to write files to the local file system and execute the included Python scripts.
Sanitization: There is no evidence of automated sanitization or filtering of the extracted content before it is used to generate skill components.
[COMMAND_EXECUTION]: The skill includes Python utility scripts (scripts/init_skill.py and scripts/package_skill.py) that perform file system operations. These operations include creating directories, writing template files, and zipping directory contents. The init_skill.py script also uses chmod to set executable permissions on generated script templates.
[DYNAMIC_EXECUTION]: A core feature of this skill is the automated generation of new executable scripts and instructions from external data. While it uses known templates for scaffolding, the generation of code from untrusted remote documentation at runtime qualifies as a dynamic execution surface.

skill-gen