Skillv1.0.14

ClawScan security

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

BenignApr 30, 2026, 4:13 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: The skill is internally consistent: it asks for a single agntdata API key and curl and its instructions are limited to calling agntdata.dev endpoints for LinkedIn-like data—no unrelated credentials, installs, or filesystem access are requested.
Guidance: This skill appears coherent, but it's a third-party aggregator (api.agntdata.dev), not LinkedIn itself. Before installing: verify agntdata's trustworthiness and terms (data sources, privacy, billing), avoid putting your primary production/LinkedIn credentials here (use a dedicated API key / account), store and rotate the AGNTDATA_API_KEY securely, and monitor API usage and billing. If you need stronger guarantees, prefer the official plugin/package referenced or test using a limited-use key first.

Purpose & Capability: okName/description claim access to LinkedIn-like data via the agntdata service; the only required capability is an AGNTDATA_API_KEY and curl, which directly align with calling that external API. Nothing requested (no other cloud creds or secrets) appears unrelated to the stated purpose.
Instruction Scope: okSKILL.md contains concrete curl examples, registration and discovery endpoints, and a list of API paths. Instructions only reference the AGNTDATA_API_KEY and standard HTTP calls to api.agntdata.dev; there are no directives to read arbitrary local files, other environment variables, or to transmit data to unexpected third parties.
Install Mechanism: okThis is instruction-only with no install spec and no code files, so nothing is written to disk or fetched during install. That is the lowest-risk install model and matches the skill's described use.
Credentials: okOnly a single environment variable (AGNTDATA_API_KEY) is required and it is declared as the primary credential — this is proportionate for an API-integration skill. No unrelated secrets, config paths, or high-privilege tokens are requested.
Persistence & Privilege: okalways is false and the skill does not request any persistent system-level privileges or modifications to other skills. The agent may invoke the skill autonomously (normal default), but that is not combined with other concerning privileges.