Molt-Solver

Automatically solves Moltbook math captchas by parsing English text, extracting operations, calculating results, and formatting answers with two decimals.

This skill appears to do what it claims (solve Moltbook math captchas and call Moltbook's API), but it reads a credentials file from ~/.config/moltbook/credentials.json and sends a POST to https://www.moltbook.com/api/v1/verify — neither of these are declared in SKILL.md or the skill metadata, and SKILL.md even claims a different dependency (memory/moltbook-state.json). Before installing or enabling this skill, consider: 1) Ask the developer to update SKILL.md and the manifest to explicitly declare the config path and required credential (what field name is expected in the file). 2) Inspect the credentials file and confirm it only contains the minimal API key the skill needs (avoid storing other secrets there). 3) If you don't trust the skill, do not provide your Moltbook API key; run it in an isolated/test account or environment first. 4) If you need stricter control, disable autonomous invocation for the agent or require manual confirmation before verification calls. 5) If you want to proceed, request the developer change the skill to read from the declared memory/moltbook-state.json or to accept the key via an explicitly-declared env/config interface so permission and auditing are clear.

Type: OpenClaw Skill Name: molt-solver Version: 1.0.0 The skill is suspicious due to its direct access and transmission of sensitive user credentials. The `molt-solver.ts` file reads `api_key` from `~/.config/moltbook/credentials.json` and transmits it to `https://www.moltbook.com/api/v1/verify`. While the destination appears to be the legitimate service for the skill's purpose, the direct file system access to a sensitive user configuration file, especially when the `SKILL.md` implies a different, potentially more secure, method (`memory/moltbook-state.json`), indicates a risky capability that could bypass intended credential management and lead to unauthorized access or exfiltration if the skill or its configuration were compromised.