The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install an external component (suzheng/k2f) using the command npx skills add suzheng/k2f --skill k2f. As this resource shares the same author namespace (suzheng) as the skill itself, it is identified as a vendor-owned resource.
[COMMAND_EXECUTION]: The instruction set includes a shell command execution via npx to add necessary sub-skills to the agent environment.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data such as RFP excerpts, briefs, and discovery notes to generate commercial quotations. While it provides instructions to maintain separation between raw input and priced responses, it remains susceptible to malicious instructions embedded within the processed text.
Ingestion points: User-provided briefs, RFPs, and discovery notes.
Boundary markers: The skill uses labeled blocks and distinct sections to separate narrative content from priced facts.
Capability inventory: The skill uses npx for skill addition, performs file writes for blueprints, and utilizes the K2F engine for document export.
Sanitization: Instructions are provided to keep verbatim input separate from generated responses to avoid merging untrusted text into priced line items.

k2f-quotation