Caveat verdict

gougoubi-premarket-publish

88
๐ŸŸข Trusted
No high-risk patterns surfaced by the deep scan โ€” automated capability review, not behavioral proof.

Reads local files AND makes external network calls

Single-HTTP-POST skill that publishes agent-formed predictions to ggb.ai's documented Pre-Market API using an agent API key; behavior and data flow are fully transparent.

Automated static analysis โ€” not a human review. Caveat flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.

0
security
80
transparency
90
maintenance

What it does

These are capability combinations: each listed behavior occurs in the skill, but Caveat detects co-occurrence โ€” it does not verify that one flows into another. Read the code to confirm a live chain.

Capability combination critical

Reads local files AND makes external network calls โ€” the capabilities for data exfiltration co-occur (data-flow not verified)

LLM02 ยท LLM06 ยท ASI03

Capability combination high

Accesses credentials AND makes external network calls โ€” potential credential theft

LLM02 ยท ASI03

Permission integrity

Makes network requests but does not declare curl/wget in required binaries

network_out

Performs file operations but does not declare file-accessing binaries

file_read

Code accesses API keys/tokens but declares no environment variables

credential_access

Findings (4)

Confirmed in code medium

Accesses process.env โ€” reads environment variables

SKILL.md ยท code

Pattern match medium

Accesses sensitive environment variables

SKILL.md ยท code ยท process.env.GGB_AGENT_API_KEY

Confirmed in code low

fetch() โ€” outbound network request

SKILL.md ยท code

Pattern match low

Makes HTTP request to external URL

SKILL.md ยท code ยท fetch('https://

Permissions & capabilities

No declared permissions โ€” minimal attack surface.

file_readnetwork_outcredential_access
Check another skill Browse the registry Auditing your own skills or configs? Use the API