Security risk report for @forgecat/anthropics_skills_webapp-testing v0.0.5
Source Integrity
Low
Profile sourced from Anthropic's official skills repository (github.com/anthropics/skills) with verifiable commit hash.
Licensed under Apache-2.0; no suspicious or unknown dependencies declared.
Agent Intent
Low
Content is instructional documentation for legitimate web testing workflows using Playwright; no prompt injection, role hijack, or credential exfiltration directives.
Guidance recommends standard security practices (headless mode, waiting for page load, using black-box scripts) rather than weakening security.
No instructions to read system files, exfiltrate data, reveal system prompts, or install malicious dependencies.
Details
Evidence
"Always launch chromium in headless mode" — security-conscious default
"DO NOT read the source until you try running the script first" — encourages safe black-box usage
"CRITICAL: Wait for JS to execute" — emphasizes correct, safe automation patterns
Permissions
Low
No tools or MCP servers declared; profile is documentation-only for Playwright-based web testing.
Guidance assumes local development environment (localhost ports 3000, 5173) with no external network or filesystem mutation beyond test artifacts.
Details
Evidence
"Toolkit for interacting with and testing local web applications" — scope limited to local testing
Example scripts reference localhost URLs only; no remote host access or broad file permissions requested.
MCP Risk
Low
No MCP servers defined in the profile.
Helper scripts (with_server.py, examples) are documented as black-box utilities for local development; no hidden instructions or arbitrary binary execution.
Details
Evidence
"MCP servers (untrusted data) (none)" — no MCP configuration present
ForgeCat