The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted job posting data via WebFetch or WebSearch to generate application responses, creating a surface for indirect prompt injection. * Ingestion points: External job postings (SKILL.md). * Boundary markers: There are no explicit delimiters or instructions for the agent to ignore instructions embedded in the job postings. * Capability inventory: The agent uses the draft_email tool to create communications (SKILL.md). * Sanitization: The skill does not implement sanitization for the content it fetches.
[DATA_EXFILTRATION]: The skill reads from local paths containing personal user data, such as /Users/forni/Craft/vocation/network/contacts.md. This access is functional for the skill's purpose but presents a data exposure surface if the agent is manipulated by malicious instructions in the external content it processes.

job-apply