The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes project documentation and source code to generate marketing plans, which creates a surface for indirect prompt injection from potentially untrusted files within a repository.
Ingestion points: The skill reads external data from docs/prd.md, README, CLAUDE.md, and package.json in SKILL.md.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to distinguish between the skill's operational instructions and the content of the project files being read.
Capability inventory: The skill utilizes several powerful tools including Write (to create documentation), AskUserQuestion, and a suite of mcp__solograph search tools for project analysis.
Sanitization: No input validation or filtering is applied to the content ingested from the project files before it is processed by the AI to generate scripts and configurations.
Remediation: Developers should wrap external file content in clear delimiters and add instructions to the agent to treat interpolated project data as passive information rather than executable instructions.

solo-video-promo