The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it fetches and follows instructions from an external documentation site and processes security findings that could contain adversarial content.
Ingestion points: docs.hacktron.ai/llms.txt (remote documentation) and api.hacktron.ai/v1/findings (finding data including proof-of-concept and descriptions) mentioned in SKILL.md.
Boundary markers: Absent. The agent is instructed to "follow the docs" and "triage each finding" without explicit delimiters or instructions to disregard embedded prompts within the external data.
Capability inventory: Across the SKILL.md workflow, the agent has access to Read, Grep, and Glob for file inspection, Write for applying fixes, and Bash for executing curl, jq, and git commands.
Sanitization: Absent. No filtering or sanitization of the remote documentation or finding data is performed before the agent processes the content.

finding-triage