The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it is designed to ingest and process untrusted external research data and direct user inputs.
Ingestion points: File SKILL.md specifies primary and secondary input sources including prioritized opportunities, direct user inputs, interview snapshots, and research findings.
Boundary markers: The skill instructions do not define explicit delimiters or 'ignore embedded instructions' markers to separate untrusted data from the core logic during topic extraction and solution generation.
Capability inventory: The skill performs file system read operations from the 'opportunities/' directory and write operations to create versioned markdown files in the 'solutions/' directory.
Sanitization: The skill mandates 'kebab-case' formatting and thematic extraction for filenames, which offers partial protection against directory traversal, but lacks comprehensive sanitization for the data interpolated into the generated solution documents.
[NO_CODE]: The skill consists exclusively of Markdown instructions for the AI agent and does not include any executable scripts, binary files, or external package dependencies.

generate-solutions