The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill describes and utilizes a data ingestion mechanism for importing previous assessment findings that lacks sufficient safety boundaries.
Ingestion points: The importAssessment logic in SKILL.md accepts external JSON data strings representing checklist items, statuses, and notes.
Boundary markers: There are no explicit delimiters or instructions to the agent to ignore potentially malicious content embedded within the imported notes or description fields.
Capability inventory: The skill is designed to search (searchChecklist), summarize, and export findings, providing multiple surfaces where an injected prompt could influence agent behavior.
Sanitization: The provided code snippets demonstrate direct parsing of JSON and storage in localStorage without sanitization or validation of the input content.
[EXTERNAL_DOWNLOADS]: The documentation references an external code repository for the tool's source code which is not from a recognized trusted vendor.
Evidence: The skill points to https://github.com/m14r41/PentestingChecklist.git for cloning the project.
[COMMAND_EXECUTION]: The skill provides explicit instructions for executing shell commands to install and run the security tool.
Evidence: Instructions include git clone, npm install, and npm run dev in the SKILL.md file.

pentesting-checklist-security-assessment