Back to skill
Skillv1.0.1
ClawScan security
ZeeLin Deep Research 深度研究 · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
ReviewMar 6, 2026, 8:44 AM
- Verdict
- Review
- Confidence
- medium
- Model
- gpt-5-mini
- Summary
- The skill appears to be a straightforward wrapper around desearch.zeelin.cn, but the runtime instructions expect an API key and use utilities not declared in the metadata, so the package metadata and the instructions are inconsistent and require clarification before trusting it.
- Guidance
- This skill appears to call the ZeeLin API and is generally coherent, but there are important mismatches you should resolve before installing: the SKILL.md expects an API key (DESEARCH_API_KEY) and uses jq in shell examples, yet the package metadata does not declare any required environment variables or jq as a required binary. Confirm the skill's source/trustworthiness (the homepage is missing), and verify the domain (desearch.zeelin.cn) is legitimate. If you plan to use it, provide the API key only when necessary (prefer per-session credentials rather than permanently exporting secrets), and be aware the skill will download report files and may send them via whatever 'message' tool your agent provides. Ask the author to: (1) declare DESEARCH_API_KEY in requires.env/primaryEnv, (2) list jq if examples require it (or remove jq from examples), and (3) provide a homepage or publisher identity. If you cannot validate the publisher, avoid installing or restrict the skill's network/file permissions and monitor outgoing requests.
Review Dimensions
- Purpose & Capability
- noteThe name/description claim a research/reporting service and the SKILL.md contains direct curl examples against https://desearch.zeelin.cn which is coherent with that purpose. Requesting curl as a required binary is reasonable. However, the SKILL.md expects an environment variable (DESEARCH_API_KEY) and uses jq in examples, but the registry metadata declares no required env vars and only lists curl; this mismatch is unexpected.
- Instruction Scope
- noteInstructions are narrowly scoped to creating/conversing with tasks on desearch.zeelin.cn, polling status, downloading report URLs, and sending files to a user. That matches the described purpose. Points to note: examples assume an environment variable DESEARCH_API_KEY and demonstrate shell snippets that use jq and a 'message' tool (message action=send ...) which are not defined in the metadata — the agent/runtime integration must provide those. The instructions also direct downloading remote PDF/asset URLs to local disk and sending them onward; this is expected for a report-download flow but is an I/O action the integrator should approve.
- Install Mechanism
- okThis is an instruction-only skill with no install spec or code to write to disk, which keeps install risk low.
- Credentials
- concernThe registry metadata lists no required environment variables, yet the SKILL.md repeatedly instructs using DESEARCH_API_KEY (x-api-key header). This is an inconsistency: the skill does expect a secret (API key) to operate but doesn't declare it in requires.env/primaryEnv. Also examples use jq without declaring it as a required binary. The skill does not request unrelated credentials (e.g., AWS), but failing to declare the API key is a proportionality/visibility problem that can hide required secret access from users.
- Persistence & Privilege
- okThe skill does not request always:true, does not claim to modify other skills or system-wide settings, and needs no special config paths. Autonomous invocation is allowed (normal), but it does not request elevated or persistent agent privileges.
