The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external data from brand assets and configuration files which could contain malicious instructions.\n
Ingestion points: Brand assets, design-system.json, and source-map.json are ingested by the agent as described in SKILL.md and references/technical.md.\n
Boundary markers: There are no explicit markers or instructions to treat ingested content as data only.\n
Capability inventory: The agent executes the astrofy CLI tool and has read access to the local file system.\n
Sanitization: Input content from external assets is not sanitized or validated before processing.\n- [COMMAND_EXECUTION]: The skill triggers the execution of a specific command-line tool.\n
Evidence: The command astrofy check --category brand is called in SKILL.md and examples/cases.md to verify the branding implementation.

astrofy-branding