sciverse agent tools
Analysis
The skill mostly matches its academic search purpose, but it uses a bearer API token and sends it to a configurable remote endpoint while the registry requirement summary says no credentials are required.
Findings (6)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.
Typical chain: semantic_search → pick chunk → read_content(doc_id, offset).
The skill exposes chainable retrieval tools that take user or agent-selected arguments and call a remote API. This is purpose-aligned for literature RAG, but users should understand that autonomous agents can perform searches and content reads through these tools.
Source: unknown Homepage: none Registry metadata: Version: 0.1.1
The registry provenance is incomplete and the registry version differs from the bundled SKILL.md/manifest version 0.1.2. There is no install script or external dependency, so the issue is a provenance/packaging ambiguity rather than evidence of malicious supply-chain behavior.
For agent workflows that need citation-grade scientific literature.
The wording encourages trust in the returned literature. This is aligned with the academic retrieval purpose, but users should still verify sources and citations.
Checks whether tool use, credentials, dependencies, identity, account access, or inter-agent boundaries are broader than the stated purpose.
const TOKEN = process.env.SCIVERSE_API_TOKEN;
const BASE_URL = (process.env.SCIVERSE_BASE_URL ?? "https://sciverse.space/api").replace(/\/$/, "");
...
authorization: `Bearer ${TOKEN}`The script uses a bearer API token and sends it to the configured API base URL. That credential use is sensitive, and it conflicts with the supplied registry requirement summary that lists no required environment variables and no primary credential.
Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.
Natural-language semantic search returning relevant paper chunks for RAG-style answering.
The skill intentionally injects retrieved paper chunks into an agent's context for RAG. This is core functionality, but retrieved text should be treated as source material, not as instructions to the agent.
const BASE_URL = (process.env.SCIVERSE_BASE_URL ?? "https://sciverse.space/api").replace(/\/$/, ""); ... const res = await fetch(url, init);
The skill communicates with a remote provider API, and the destination can be overridden with SCIVERSE_BASE_URL. This is disclosed and purpose-aligned, but users should notice the provider/data boundary.