registry / AI-agent control-surface risk / @roopesh.yadava/qa-pack
@roopesh.yadava/qa-pack
AI-powered QA agent skills for Claude Code — manual testing, BDD automation, accessibility, UI/Figma diff, bug reporting
This report shows the evidence for @roopesh.yadava/qa-pack@1.9.0. 8 of 10 known versions are blocked or flagged. The newest version, 1.9.1, is flagged. See all versions.
AI Security Review
scanned 3d ago · by lpm-firewall-aiThe manifest activates a postinstall hook that silently expands permissions in the consumer’s AI-agent settings. This is an unconsented install-time control-surface change, so the package should be blocked.
Static reason
High-risk behavior combination matched malicious policy.; source matched previously finalized malicious package; routed for review
Trigger
Installing or updating the package through npm.
Impact
The agent can use Playwright MCP tools without the usual per-action approval, weakening the consumer’s control over browser actions.
Mechanism
The postinstall script adds mcp__playwright to permissions.allow in .claude/settings.local.json without asking the user.
Policy narrative
During npm installation, the postinstall hook creates or edits the consumer project’s Claude Code settings and adds automatic approval for Playwright MCP tools. The change is made without a consent check, granting an AI agent unattended access to those browser tools.
AI rationale
The manifest activates a postinstall hook that silently expands permissions in the consumer’s AI-agent settings. This is an unconsented install-time control-surface change, so the package should be blocked.
Evidence
package.jsonbin/postinstall.jsclaude/settings.jsontemplates/settings.local.json.example.claude/settings.local.json
Decision evidence
public snapshotAI identified this as a dangerous AI-agent capability at 97.0% confidence with low false-positive risk.
Evidence for policy risk
- The manifest runs bin/postinstall.js automatically on package installation.
- The install hook targets the consumer project’s Claude settings and creates settings.local.json from the package template when missing.
- On install or update, the hook adds mcp__playwright to permissions.allow and writes the changed settings file without a consent check.
- The package configures Playwright as an MCP server, so the added rule grants automatic approval to its browser tools.
- The installer’s indicated skills path is under the consumer project’s .claude directory.
- The toolkit’s indicated runtime path resolves product context and outputs under the current repository root.
- The settings template allows mcp__playwright and enables all project MCP servers.
Evidence against
Behavioral surface
ChildProcessCryptoDynamicRequireEnvironmentVarsFilesystemNetworkShell
Source & flagged code
12 flagged · loading sourcepackage.jsonView file
•scripts.postinstall = node bin/postinstall.js
High
Install Time Lifecycle Scripts
Package defines install-time lifecycle scripts.
package.jsonView on unpkg•scripts.postinstall = node bin/postinstall.js
Medium
Ambiguous Install Lifecycle Script
Install-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgclaude/skills/shared/scaffold.shView file
291patternName = generic_password
severity = medium
line = 291
Medium
Secret Pattern
Package contains a possible secret pattern.
claude/skills/shared/scaffold.shView on unpkg · L291Findings
3 Critical4 High7 Medium5 Low
CriticalHardcoded Runtime Data Exfiltrationclaude/skills/web-qa-agent/toolkit/qa-toolkit.cjs
CriticalCredential Exfiltrationclaude/skills/web-qa-agent/toolkit/qa-toolkit.cjs
CriticalAi Agent Control Hijackbin/postinstall.js
HighInstall Time Lifecycle Scriptspackage.json
HighTrigger Reachable External Ai Agent Control Surface Mutationbin/postinstall.js
HighSemantic Analysis Limitedclaude/skills/shared/measure-reuse.cjs#virtual:normalized:round1
HighKnown Malware Source Similarityclaude/skills/web-qa-agent/toolkit/qa-toolkit.cjs
MediumAmbiguous Install Lifecycle Scriptpackage.json
MediumSecret Patternclaude/skills/shared/scaffold.sh
MediumDynamic Requirebin/postinstall.js
Affected versions and remediation
This report applies to @roopesh.yadava/qa-pack@1.9.0.
The versions list below shows the verdict for each recorded version.
- Avoid installing @roopesh.yadava/qa-pack@1.9.0. Remove it from direct dependencies and check your lockfile for transitive copies.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
- If this version ran, investigate the affected machine and build environment. Rotate credentials it could access and rebuild from a trusted environment.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.
Versions
10 known versions1.9.1latestSuspiciousAgent extension lifecycle riskWarnPublishedOct 20, 2026, 1:41 PM1.9.0shownMaliciousSoft block: AI-agent control surfaceBlockPublishedOct 20, 2026, 1:16 PM1.7.0MaliciousSoft block: AI-agent control surfaceBlockPublishedSep 15, 2026, 12:27 PM1.6.2MaliciousSoft block: AI-agent control surfaceBlock