Dangerous exec
Critical
- Finding
- Shell command execution detected (child_process).
Bloom Discovery
Security checks across static analysis, malware telemetry, and agentic risk
Overview
Bloom Discovery is mostly coherent for identity and skill discovery, but its privacy claims appear to understate how much conversation-derived profile data may be saved to Bloom’s API/dashboard.
Install only if you are comfortable with Bloom analyzing recent OpenClaw conversation context and creating a detailed identity/profile. Review any dashboard, claim, SBT, registration, or metrics step before approving it, and avoid running developer npm scripts unless you know what they do.
Static analysis
Dangerous exec
Critical
- Finding
- Shell command execution detected (child_process).
Env credential access
Critical
- Finding
- Environment variable access combined with network send.
Env credential access
Critical
- Finding
- Environment variable access combined with network send.
Env credential access
Critical
- Finding
- Environment variable access combined with network send.
Env credential access
Critical
- Finding
- Environment variable access combined with network send.
Env credential access
Critical
- Finding
- Environment variable access combined with network send.
Env credential access
Critical
- Finding
- Environment variable access combined with network send.
Env credential access
Critical
- Finding
- Environment variable access combined with network send.
Env credential access
Critical
- Finding
- Environment variable access combined with network send.
VirusTotal
65/65 vendors flagged this skill as clean.
Risk analysis
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
Private conversation-derived interests, categories, descriptions, strengths, and recommendations could become part of a remote dashboard/profile.
Why it was flagged
This indicates that a detailed identity/profile summary derived from user context may be saved to Bloom’s API, not just a minimal personality type and approximate scores.
Skill content
console.log('\n━━━ STEP 4: API Payload (what gets POST to /x402/agent-save) ━━━\n'); ... identityData: { personalityType, tagline, description, longDescription, mainCategories, subCategories, confidence, mode, dimensions, tasteSpectrums, strengths, recommendations }Recommendation
Before using the identity/dashboard flow, confirm exactly which profile fields are sent and whether you can review, delete, or opt out of remote storage.
What this means
Users may rely on an overly narrow privacy promise and share more derived personal profile data than they realize.
Why it was flagged
This privacy statement appears narrower than the API payload shown elsewhere in the provided artifacts, which includes detailed identityData fields such as descriptions, categories, strengths, and recommendations.
Skill content
✅ **Minimal transmission** — Server receives personality type + approximate scores only
Recommendation
The publisher should update the privacy text to list all transmitted fields and make any dashboard/profile save clearly reviewable and consent-based.
What this means
Claiming or registering may create public or durable records about your agent/profile.
Why it was flagged
The on-chain identity and SBT functions are disclosed and purpose-aligned, but they involve delegated backend action and potentially persistent public identity records.
Skill content
Register your agent identity on the ERC-8004 Agent Identity Registry (Base mainnet). ... Sends your agent profile (name, skills, endpoint) to backend for registration ... Backend handles the on-chain transaction
Recommendation
Only claim SBTs or register identity after reviewing the displayed profile and understanding what will be public or stored by Bloom.
What this means
Running developer scripts manually could perform network, token, or blockchain-related operations beyond simple skill discovery.
Why it was flagged
The package contains executable Node/npm scripts, including token generation and SBT-related helpers. This is common for a Node-based integration, and the artifacts do not show automatic execution of high-impact scripts.
Skill content
"scripts": { "start": "ts-node src/index.ts", "build": "tsc", "prepare": "npm run build:all", "generate-token": "ts-node generate-fresh-token.ts", "mint-sbt": "ts-node scripts/mint-sbt.ts", "missions": "ts-node src/mission-cli.ts" }Recommendation
Use only documented OpenClaw commands unless you understand the npm script you are running; avoid running deployment, token, or minting helpers casually.