pp-learn-loop-example

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs using a CLI subcommand that embeds a token as a command-line argument (learn-loop-example-pp-cli auth set-token YOUR_TOKEN_HERE), which would require the LLM/agent to handle or emit secret values verbatim and thus creates an exfiltration risk.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). These are GitHub-hosted project pages; the specific GitHub release URL (printing-press-library/releases/...) points to pre-built installer/bundle assets — downloading executables from personal or project GitHub releases is higher-risk than using official vendor channels.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow for this skill’s learn loop ingests the outsider-authored free-text query parameter (e.g., the user’s question passed to learn-loop-example-pp-cli recall "<question>" --agent / teach --query "<question>" ...) and then feeds that content back into the LLM-consumed JSON envelope and learning/journal state without requiring selection of a specific trusted item.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The README instructs users/agents to download and run a pre-built binary from https://github.com/mvanhorn/printing-press-library/releases/tag/learn-loop-example-current (and the SKILL.md recommends installing the CLI via npx), which are runtime steps that fetch and execute remote code and the skill requires that CLI to be present.