Trawl

Autonomous lead generation through agent social networks. Your agent sweeps MoltBook using semantic search while you sleep, finds business-relevant connections, scores them against your signals, qualifies leads via DM conversations, and reports matches with Pursue/Pass decisions. Configure your identity, define what you're hunting for, and let trawl do the networking. Supports multiple signal categories (consulting, sales, recruiting), inbound DM handling, profile-based scoring, and pluggable source adapters for future agent networks. Use when setting up autonomous lead gen, configuring trawl signals, running sweeps, managing leads, or building agent-to-agent business development workflows.

This skill appears to do what it says: it searches MoltBook, scores profiles, opens/approves DMs, and stores leads locally. Before installing, consider: 1) Ensure you trust the MoltBook API and supply only the MOLTBOOK_API_KEY (keep it in ~/.clawdbot/secrets.env as instructed). 2) Confirm you have the required CLI tools (curl, jq, bc, column and standard date utilities) or the scripts will fail — the metadata does not declare these dependencies. 3) Review config.json especially auto_approve_inbound (defaults to false) to avoid auto-accepting inbound DMs unintentionally. 4) The skill writes state to ~/.config/trawl and reads ~/.clawdbot/secrets.env — verify those paths and the files before running. 5) The source is listed as unknown and there's no homepage; if provenance matters, prefer packages with a known author or repository. If you want higher confidence, ask the publisher for a canonical repo or signed release and/or run the scripts in a disposable environment first.

Type: OpenClaw Skill Name: trawl Version: 1.0.2 The skill is classified as suspicious due to `jq` injection vulnerabilities found in `scripts/leads.sh` and `scripts/report.sh`. In both scripts, user-controlled filter variables (`STATE_FILTER`, `CAT_FILTER`) are directly interpolated into `jq` filter expressions without proper escaping (e.g., `select(.value.state == "$STATE_FILTER")`). This flaw could allow an attacker to inject arbitrary `jq` syntax, potentially leading to unauthorized disclosure or manipulation of data within the local `leads.json` or `last-sweep-report.json` files. While the external API interactions are handled more securely with URL encoding and safe JSON construction, this internal data processing vulnerability is a significant concern.