ClawHub
Back to skill

Security audit

Phy Ai Context Gen

Security checks for vulnerabilities and agentic risk

Overview

This skill transparently scans project files to generate AI assistant instruction files, with some secret-handling caveats users should review.

Install only if you want a skill that reads project metadata and creates assistant instruction files. Before committing or reusing the generated files, check that they do not include secret values, private internal details, or overly broad agent permissions such as unrestricted repository reads.

A.I.G

Vulnerability Patterns
  • 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
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Credential Access

High
Category
Privilege Escalation
Confidence
98% confidence
Finding

The generated CLAUDE.md template explicitly calls out .env files in "Important Notes," and earlier the skill says it generates project instructions including "API keys," which normalizes collecting or documenting secret locations for downstream agents. Even if it does not directly exfiltrate credentials, encouraging inclusion of secret-bearing files in assistant context increases the likelihood that future prompts or tools will read, summarize, or leak sensitive material.

Content

Scanner excerpt · SKILL.md (reported line 162)May include surrounding context.

text

## Important Notes
- [any .env files to be aware of]
- [any generated files not to edit]
- [any protected paths or files]

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger phrases are broad, natural-language requests like "set up AI context" and "help the AI understand this project," which can match ordinary user intent and invoke the skill unexpectedly. Unintended invocation matters here because the skill proceeds to read many repository files and generate/write multiple outputs, increasing the chance of surprise data access or side effects.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill instructs broad repository inspection via commands like find, cat package.json, cat README.md, config-file reads, and reading existing context files, but provides no warning or guardrails around secrets and sensitive configuration. In real repositories, these patterns can easily surface tokens, internal URLs, credentials, or security-relevant settings from manifests, CI files, and adjacent configs, exposing more information than the user may expect.

Content

No source excerpt is available for this finding.