The Agent Skills Directory

[SAFE]: The skill implements a controlled workflow for reading and writing documentation files. It operates within a defined project workspace, accessing expected upstream documents and writing to specific stage outputs like 06-feasibility.md and status.json.
[EXTERNAL_DOWNLOADS]: The skill utilizes the agent's web search capability to verify technical requirements, costs, and market data as outlined in references/research-policy.md. This usage is consistent with the skill's primary purpose and does not involve downloading or executing untrusted code.
[PROMPT_INJECTION]: The skill reads data from upstream files (03-scenarios.md, 04-painpoints.md, 05-value.md) which may contain content derived from previous user input, creating a surface for indirect prompt injection. This is evaluated as safe because the skill's capabilities are limited to summarization and assessment.
Ingestion points: Upstream documentation files listed in references/stage-contract.md.
Boundary markers: None present in the provided markdown templates.
Capability inventory: Local file read/write access and Web Search tool usage.
Sanitization: No specific input sanitization or filtering logic is defined for the input data.

feasibility-assess