Wyld Stallyns: Be Excellent 🎸

Summon legends into the booth. 14 philosophers, warriors, artists, leaders to help with decisions, creative work, and life's hard questions. Marcus Aurelius for when you're spiraling. Bruce Lee for when you're too rigid. Tubman for when you're scared. Munger for when you're fooling yourself. Or forge your own with Rufus as your guide. Be excellent to each other. 🎸

This skill appears coherent and self-contained: it uses the included legend files to produce perspective-based responses and does not require network access, credentials, or installs. Two practical checks before enabling it: (1) Decide whether you want to allow the 'Forge' capability — it needs the agent runtime to be able to write files in the skill directory, so confirm write permissions are restricted to this skill and can't alter other skills or system files. (2) If you have concerns about content (e.g., how real historical figures are represented), review the bundled legend markdowns; the skill will present those texts as guidance. If you don't want any on-disk persistence, disable or avoid using summon forge/retire operations.

Type: OpenClaw Skill Name: wyld-stallyns Version: 1.0.4 The skill is suspicious due to the `summon forge <candidate>` command described in `SKILL.md` and detailed in `FORGE.md`. The `FORGE.md` instructions direct the AI agent to 'research the candidate deeply' (potential for external interaction/search query injection) and, more critically, to 'Save markdown file: `skills/wyld-stallyns/assets/legends/[legend-id].md`' and 'Add to `council.json`'. These file write operations, based on user-provided input for `<candidate>`, present a significant risk for path traversal (writing files outside the skill's directory) and prompt injection (injecting malicious markdown instructions into the generated `.md` file or malforming `council.json`), as no input sanitization is explicitly mentioned.