Versioned security assessment

Report ID: SA-EBDFE608

7/23/2026, 6:44:15 PM

flow-nexus-platform security assessment v8

Skill Security Certification Report

Audit History
Scanner version 3.0.0 Audit model: codex Latest published report
Skill name
flow-nexus-platform
Version
v8
Maintainer
ruvnet
Source
github.com/ruvnet/claude-flow/tree/main/.claude/skills/flow-nexus-platform
Coverage
1 Files scanned · 1,155 Lines analyzed
Policy version
skillstore-security-audit-policy-v1

Highest confirmed finding severity

High

6 confirmed security findings require attention.

Installation context

Check the current Skill page

This page summarizes report evidence only. The Skill page provides the canonical install advisory.

Open current Skill page

This report does not block or authorize the manifest or ZIP.

Most static alerts are false positives caused by Markdown fences, JavaScript template strings, and placeholder credentials. Confirmed risks include outbound network access, repository cloning, remote command execution, secret transmission, paid actions, delegated tools, and destructive platform mutations. No prompt injection, obfuscation, or embedded live credential was found.

Report position

Latest published report

Latest refers to the report sequence, not to artifact currentness.

Audit attestation

Active attestation

A public attestation is available for this exact report.

Human verification

Not verified

No human verification is recorded for this report.

Coverage

1 Files scanned · 1,155 Lines analyzed

10 items shown for review

Limitations

This report does not claim runtime or sandbox execution and does not prove the absence of side effects.

Evidence chain

Follow the evidence from source binding to the install contract. Available evidence supports verification; it is not a safety guarantee.

  1. Source

    Commit and path bound

  2. Artifact

    Content and tree hashes bound

  3. Audit

    Complete

  4. Install contract

    Open manifest to verify

    Open manifest

Capabilities observed

Observed means this report recorded supporting evidence. Not recorded does not prove that a capability is absent.

Contains scripts

May execute code included with the Skill.

Not recorded by this audit

Network access

May connect to external services.

Observed in 11 evidence locations

Filesystem access

May read or write local files.

Not recorded by this audit

Env variables

May read values from the process environment.

Observed in 10 evidence locations

External commands

May invoke commands or programs outside the Skill.

Observed in 51 evidence locations

Capability review items (4)
Medium
SKILL.md:1053-1055
Ruby/shell backtick execution
`
The backtick closes a startup script containing apt-get, git clone, and npm install commands sent to a remote sandbox. Following the example executes commands and installs third-party code.
Medium
SKILL.md:1051
Hardcoded URL
git clone https://github.com/user/repo
The startup script clones a remote repository and then runs npm install. A mutable or untrusted repository can introduce malicious package code and install hooks.
Low
SKILL.md:162
Fetch API call
const result = await fetch('https://api.example.com/data');
The example passes a fetch call to sandbox_execute, causing a real outbound request when followed. The endpoint is a placeholder, but remote network access is explicit.
Low
SKILL.md:162
Hardcoded URL
const result = await fetch('https://api.example.com/data');
The example passes a fetch call to sandbox_execute, causing a real outbound request when followed. The endpoint is a placeholder, but remote network access is explicit.

Risk findings

Confirmed security concerns are separated from items that still need review.

Confirmed security concerns (6)

RISK-001 High
SKILL.md:141-174SKILL.md:1043-1071
Remote Code and Command Execution
The skill instructs agents to install packages and execute caller-supplied code and shell commands in remote sandboxes. Malicious inputs or dependencies can execute untrusted code and consume resources.
The cited examples directly provide run_commands, startup_script, package installation, and sandbox_execute inputs.
RISK-002 High
SKILL.md:121-174SKILL.md:274-290SKILL.md:378-394
Sensitive Credentials Sent to Hosted Operations
Examples pass API keys, database URLs, application secrets, and an Anthropic key into Flow Nexus operations. Real replacements would disclose credentials to a community-operated external service.
The documented MCP parameters explicitly accept API keys, database connection strings, secrets, and an Anthropic key.
RISK-003 High
SKILL.md:480-506SKILL.md:886-903
Financial Actions and Recurring Charges
The skill can create payment links and enable automatic credit purchases. Agent execution without explicit approval could cause unintended charges.
The examples directly configure a payment amount and enable auto-refill with a recurring purchase threshold.
RISK-004 High
SKILL.md:819-836
Delegated Assistant Can Invoke Tools
The Seraphina example enables tools that can create swarms and deploy code. Delegation can expand impact beyond the user's immediate request.
The example sets enable_tools to true and states that the assistant can create swarms and deploy code.
RISK-005 High
SKILL.md:209-220SKILL.md:342-394SKILL.md:740-746
Destructive and Public Platform Mutations
The skill documents sandbox deletion, storage deletion, application publication, application updates, and template deployment. Incorrect targets can destroy data or expose unreviewed source code.
The cited MCP calls directly delete resources, publish source code, update applications, and deploy templates.
RISK-006 High
SKILL.md:107-113SKILL.md:528-536
Caller-Supplied Entitlement and Credit Mutations
Examples accept caller-supplied user identifiers, tiers, credit amounts, and reasons for entitlement changes. Weak server authorization could permit balance or tier abuse.
The parameters support direct account and credit mutations, but server-side authorization behavior is not present in the audited file.

Remediation

Suggested fixes recorded by this audit. Applying them is the maintainer’s responsibility.

  1. FIX-001
    High
    Remote code, shell commands, and packages can execute untrusted content.
    Require a user-approved command and package plan, pin trusted versions and commits, disable package lifecycle scripts where possible, and enforce sandbox resource limits.
  2. FIX-002
    High
    Examples encourage sending API keys and database credentials to hosted operations.
    Use scoped secret references from an approved secret store, redact logs, prevent plaintext prompt entry, and document the external trust boundary.
  3. FIX-003
    High
    Payment links and auto-refill settings can create charges.
    Require explicit confirmation that shows amount, currency, threshold, recurrence, and account before every paid or auto-refill change.
  4. FIX-004
    High
    Publishing, deployment, account changes, and deletions can create irreversible effects.
    Add previews, ownership checks, scoped authorization, target identifiers, and separate confirmations for public, destructive, or entitlement-changing operations.
  5. FIX-005
    High
    Delegated assistant tools can create swarms and deploy code.
    Default delegated tools to disabled, use a minimal allowlist, constrain budgets and destinations, and require approval for each external action.

Expert evidence

Immutable subject identity, scanner metadata, dismissed matches, and source-level evidence.

Artifact subject

Marketplace commit
ebdfe608f5de2b66ff37ab4af12af8ac4f5e8006
Content hash
7b0b002fa0edda7130a095c1df192c4f9ffc7aaf59936d28ad5de22212eb6486
Tree hash
6ecacc271ddd96e739cfd55fc24056fdb75db49749a0c39c97b446021319beed
Skill path
skills/ruvnet/flow-nexus-platform
Audit payload hash
36c62ab57d6cc1fccf72a47cf1e222e6

Analysis metadata

Audit model: codex

Analysis state: Complete

Scope is limited to the recorded files, lines, methods, and evidence. No runtime or sandbox execution is claimed.

Verify and export

JSON report Signed attestation CITATION.cff SARIF 2.1.0 Signed installation manifest Source

The manifest and lockfile bind install artifacts to cryptographic hashes. This integrity claim is separate from the security assessment.

Audit attestation: active