The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes external data that could potentially contain malicious content intended to influence agent behavior.
Ingestion points: Picoclaw configuration files (e.g., config.json, config.yaml, launcher-config.json) are analyzed by lib/profile.mjs, and security advisory feeds are processed by lib/advisories.mjs.
Boundary markers: Absent. The skill does not utilize specific delimiters or instructions to isolate untrusted data from processing logic.
Capability inventory: The skill can perform file system operations (read and write) through its associated Node.js scripts.
Sanitization: Employs JSON.parse and regular expressions for analysis but does not implement advanced sanitization for prompt injection within the data context.
[COMMAND_EXECUTION]: The documentation provides shell command blocks for manual release artifact verification.
Evidence: SKILL.md contains a shell script using curl, openssl, jq, and shasum to verify the integrity and provenance of downloaded components.
[EXTERNAL_DOWNLOADS]: The skill references external sources for installation and verification purposes.
Evidence: Fetches release artifacts, checksum manifests, and signing keys from the official GitHub repository github.com/prompt-security/clawsec.

picoclaw-security-guardian