Mastodon Scout

Read-only Mastodon skill. Outputs human-readable timeline summaries or raw JSON.

This skill appears to do exactly what it claims: a read-only Mastodon timeline/search helper implemented in a small Python script. Before installing: (1) confirm the registry metadata is updated to declare MASTODON_TOKEN (and optionally MASTODON_INSTANCE/LIMIT) so you know what will be required; (2) create a Mastodon application token with only the read scope (the SKILL.md explicitly recommends this) and keep the token secret — do not paste it into public logs or version control; (3) verify the Mastodon instance URL you provide is correct (the script will call whatever instance you pass); and (4) if you allow autonomous invocation of skills, remember this one can run with your token when invoked, so avoid granting it overly broad tokens. If you want extra assurance, inspect the included scripts/mastodon_scout.py locally before running it — it uses only stdlib network calls and only performs GET requests to standard Mastodon API endpoints.

Type: OpenClaw Skill Name: mastodon-scout Version: 2.0.1 The OpenClaw skill 'mastodon-scout' is designed for read-only interaction with the Mastodon API. The `SKILL.md` provides clear, security-conscious instructions for token setup, explicitly limiting scopes to 'read'. The `scripts/mastodon_scout.py` uses standard Python libraries, handles user input (e.g., search queries) with `urllib.parse.quote` to prevent injection, and only makes GET requests to the configured Mastodon instance. There is no evidence of data exfiltration, malicious execution, persistence mechanisms, or prompt injection attempts against the agent beyond its stated purpose of providing user guidance for authentication.