Skillv1.0.4

ClawScan security

last30days · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignApr 13, 2026, 11:40 PM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's code, declared requirements, and runtime instructions are coherent with a social/web aggregation research tool; requested credentials and file access align with its stated purpose.
Guidance: This skill appears to do what it says: aggregate recent social and web signals and store/summarize them. Before installing or running it, consider: (1) AISA_API_KEY will be sent to the AISA service — use a scoped/test key if you have concerns about sending data to that provider. (2) The skill writes a local SQLite DB and briefing files (by default under the repo and ~/.local/share); if you prefer isolation, run it in a dedicated directory or container. (3) Optional GitHub tokens may be requested only for GitHub lookups — supply them only if you need that feature. (4) The setup flows can persist config and perform device-auth; read scripts/lib/env.py and scripts/lib/aisa.py (or search for where credentials are written) if you want to audit exact persistence behavior. (5) Because the skill can run code and make network requests, run it in an environment you control (or review the bundled source) if you are concerned about data leakage.

Review Dimensions

Purpose & Capability: okName/description (aggregate last-30-days social + web sources) match the included code and scripts. Declared binaries (python3, bash) and primary env var (AISA_API_KEY) are expected because the implementation uses an AISA-hosted planner/provider for many backends; GitHub token is optional and documented.
Instruction Scope: noteSKILL.md and scripts instruct the agent to run local Python scripts that fetch, score, cluster, persist, and render results. The skill reads/writes a local SQLite store and saves briefings under user paths (~/.local/share/last30days and ./ .last30days-data). It also exposes setup flows that may perform device-auth or write configuration. These behaviors are within the scope of a research aggregator but are persistent and should be expected by the user.
Install Mechanism: okNo install spec (instruction-only + shipped scripts). No remote downloads or archive extraction. The runtime is pure local Python/bash execution — low install risk. Review of included files is possible because source is bundled.
Credentials: okOnly one required env var is declared (AISA_API_KEY) and is used for AISA planner/provider calls. GH_TOKEN/GITHUB_TOKEN are optional and documented for GitHub API use. No unrelated secrets are demanded.
Persistence & Privilege: noteThe skill persists data to local disk (SQLite DB under ./ .last30days-data and briefings under the user's home), can write setup config via the setup wizard, and may create tokens via device-auth flows. always:false; it does not force inclusion. Persistence and ability to store credentials are expected but worth noting.