Tootbot

Publish content to Mastodon. Use when you need to post a Mastodon status.

Do not install or run this skill blindly. Specific things to check before using it: - Verify the source and publisher: the registry metadata lists no homepage and the owner id is unfamiliar. - Expect to provide MASTODON_URL and MASTODON_ACCESS_TOKEN; only do so if you trust the skill and have verified the script. - Inspect the full scripts/tootbot.js in a safe environment (or ask the author for readable source). The file is large and minified — consider requesting an unminified repository reference or source review. - Validate the access token's scope (create a token with minimal scope needed for posting) and consider using a throwaway/test account first. - Run the script in an isolated container or sandbox and monitor network calls (to confirm it only talks to the configured Mastodon instance). - If you cannot review the code or verify provenance, treat this skill as untrusted and avoid supplying your real Mastodon access token.

Type: OpenClaw Skill Name: tootbot Version: 0.5.0 The skill is classified as suspicious due to two main factors: 1) The `scripts/tootbot.js` file allows local file access for media uploads (e.g., `"file": "/path/to/foo.png"`). While necessary for its stated purpose, this capability could be exploited if the AI agent is prompted to provide a sensitive file path. 2) The `SKILL.md` contains instructions like `Read the output and summarize it for the user.`, which represents a generic prompt injection vector. Although the script's output is controlled JSON from the Mastodon API, a malicious API response could potentially be crafted to influence the agent's subsequent actions during summarization. There is no clear evidence of intentional harmful behavior from the skill developer, but these capabilities introduce elevated risks.