Back to skill

Security audit

Multi-Channel Skill Publisher

Security checks for vulnerabilities and agentic risk

Overview

This skill is a publishing checklist with no executable payload, but users should review files before following its public GitHub push steps.

Before using the release commands, run `git status --short`, inspect the staged diff, confirm `.gitignore` covers secrets and local files, and consider using explicit `git add` paths instead of `git add -A`, especially before creating or pushing to a public repository.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (1)

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:59
Finding

Blanket Git staging may expose sensitive files through a public repository

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 59–63
Vulnerability Type: Unsafe release and source-control procedure
Risk Level: Medium

Complete Code Snippet:

bash
git add -A && git commit -m "Version -> <v>" && git push
git rev-parse HEAD
bash
gh repo create <owner>/<repo> --public --source . --push

Technical Analysis

The documented release procedure uses git add -A, which stages every changed, deleted, or untracked file not excluded by Git ignore rules. It then instructs the user to push the resulting commit. For an initial release, it explicitly creates a public GitHub repository and pushes the current source tree.

The procedure does not require users to validate .gitignore, inspect the staged diff, or run secret scanning before publication. Consequently, an unignored sensitive file in the working tree—such as an environment file, private key, access token, internal configuration, personal data, or unrelated proprietary artifact—can be included in the commit and disclosed publicly.

This is an unsafe release practice rather than evidence of intentional exfiltration. The publishing and network operations otherwise correspond to the Skill's declared purpose.

Attack Path

  1. A sensitive file exists within the repository working tree and is not covered by .gitignore.
  2. The user follows the Skill's release procedure and executes git add -A.
  3. Git stages the sensitive file together with the intended release files.
  4. The user commits without reviewing the staged file list and diff.
  5. git push, or gh repo create ... --public --push, publishes the commit to a public GitHub repository.
  6. External parties can retrieve the exposed data from the repository or its Git history, including after a later working-tree deletion unless the history is rewritten.

Impact Assessment

The issue does not directly grant additi ...[truncated 621 chars]

Remediation
View remediation

Remediation Suggestions

  1. Replace blanket staging with an explicit allowlist of expected release files, for example:
    bash
    git add SKILL.md skill.json sync.py clawhub-skill plugin .claude-plugin
    
  2. Require review before every commit and push:
    bash
    git status --short
    git diff --cached --stat
    git diff --cached
    
  3. Establish and validate a comprehensive .gitignore covering environment files, credentials, private keys, local configuration, build output, editor files, and temporary artifacts.
  4. Run a secret scanner such as Gitleaks or TruffleHog against both the working tree and staged content before publication.
  5. Require explicit user confirmation before creating a public repository; default to a private repository when public access is not necessary.
  6. Add an automated pre-push or CI control that rejects commits containing high-confidence secrets or prohibited file patterns.
  7. Document an incident-response procedure for accidental disclosure: revoke and rotate exposed credentials immediately, remove sensitive content from Git history, force-push sanitized history where appropriate, and review downstream mirrors or registry copies.
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep