The Agent Skills Directory

[SAFE]: The skill provides structured guidance and professional templates for grant writing tasks. All content is educational and procedural, aligning perfectly with the stated role of a grant writer specialist.\n- [NO_CODE]: There are no scripts, binaries, or automated shell commands present in the skill files. The skill operates entirely through natural language instructions.\n- [INDIRECT_PROMPT_INJECTION]: The workflow involves the agent reading external funder websites, RFPs, and financial documents. This creates a potential surface for indirect prompt injection if external content contains malicious instructions. However, the risk is considered safe because the skill is restricted to text production and lacks the ability to execute code or access the local filesystem/network independently.\n
Ingestion points: Funder websites, GrantStation, Foundation Directory, and PDF/RFP documents referenced in the prospect research and proposal development workflows (SKILL.md).\n
Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands in external text.\n
Capability inventory: Text generation only; no tool access for web requests, file writing, or command execution is enabled in the configuration.\n
Sanitization: No explicit sanitization or content validation procedures are described for processing external data.

grant-writer