ClawHub
Back to skill

Security audit

paper-reader

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent paper-ingestion workflow, but it asks an agent to run a broad external CLI and automatically commit and push repository changes without a clear approval gate.

Install only if you are comfortable with the workflow modifying your paper repository and pushing to its remote. Before use, add a manual confirmation step before any Kimi invocation and before git commit/push, review `git status` and diffs, and replace `git add -A` with staging only the intended paper and index files.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Context-Inappropriate Capability

Medium
Confidence
90% confidence
Finding
The skill directs use of an external interactive Kimi CLI in `--yolo` and background/PTy mode, which weakens containment and gives a second agent broad, hard-to-audit operating latitude. Even though the stated task is document review, the invocation pattern is capable of performing actions beyond read-only analysis, increasing the chance of unintended file changes, command execution, or data exposure.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill performs `git add -A`, `git commit`, and `git push` automatically as part of the workflow without a distinct confirmation step immediately before repository mutation and remote publication. This creates a real risk of committing unintended files, sensitive content, or low-quality/generated changes and exfiltrating them to a remote repository without the user's informed consent.

Static analysis

No suspicious patterns detected.