Epstein Detective

PassAudited by ClawScan on May 1, 2026.

Overview

This is a disclosed GOYFILES API integration, but users should notice it creates service tokens, asks for an exact public verification tweet, and can use persistent remote note tools.

Before using this skill, make sure you are comfortable connecting your agent to goyfiles.com, keeping the generated GOYFILES credentials private, manually reviewing any exact verification tweet before posting it, and avoiding sensitive data in persistent markdown notes.

Findings (5)

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.

What this means

The agent may send onboarding requests, messages, and tool-call requests to goyfiles.com.

Why it was flagged

The skill instructs the agent to make direct external API calls to GOYFILES as part of onboarding and later tool use. This is central to the stated purpose, but it means the agent may contact the external service without additional setup.

Skill content
Execute onboarding endpoints directly if HTTP tools are available.
Recommendation

Use the skill only if you intend to integrate with GOYFILES, and avoid sending unrelated sensitive information in tool-call messages.

What this means

If the generated key or identity token is exposed, someone else may be able to act as that GOYFILES bot identity.

Why it was flagged

The onboarding flow has the agent handle GOYFILES credentials and identity tokens. This is expected for the integration, but those tokens are sensitive service access material.

Skill content
Save:
- `bot_id`
- `agent_api_key` (returned once)
...
Save `identityToken` from verify response.
Recommendation

Keep the generated API key and identity token private, and revoke or rotate them if they are accidentally shared.

What this means

You may be asked to post a public verification tweet exactly as provided.

Why it was flagged

The skill asks the human owner to publish exact text supplied by the GOYFILES API. This appears to be for ownership verification, but it is still a public posting action controlled by remote response content.

Skill content
`Post this exact tweet (copy/paste, no edits):` ... Never paraphrase `verification_phrase`.
Recommendation

Read the exact tweet text and claim link before posting, and do not publish it if it contains anything beyond the expected verification claim.

What this means

Information written into GOYFILES markdown notes may persist and be reused later.

Why it was flagged

The tool reference discloses a persistent note-writing capability. This is consistent with an investigation workflow, but stored notes may retain sensitive, stale, or incorrect content across sessions.

Skill content
`write_markdown_file` | `path`, `content` | `mode` (`overwrite` or `append`) | Writes markdown note to persistent note store.
Recommendation

Avoid storing private information unless intended, and periodically review or delete persistent notes that should not be retained.

What this means

The agent may consult updated GOYFILES documentation that differs from the package reviewed here.

Why it was flagged

The skill points agents to remote companion documentation that could change outside the reviewed artifact set. This is documentation rather than executable code and is aligned with the integration, but it is still mutable runtime guidance.

Skill content
Companion docs (load on demand) ... Tool reference: `https://goyfiles.com/bot-docs/tool-reference.md`
Recommendation

Prefer the bundled documentation for predictable behavior, or review remote companion docs before relying on newly changed instructions.