SECURITY WORKFLOW COMPONENTS / Anthropic
Claude Security patch verifier
A security-patch verifier that derives the staged diff and separates tested behavior from source-only confidence.
“Claude Security component for independently examining a staged security fix in a scratch workspace.”
01 / THE REASONING
Why this made the selection.
- It rejects unstaged/untracked changes, derives the diff with external drivers disabled and checks the pre-change exploit before judging a fix.
- The behavior claim carries untested=true when no repository test exercises the changed code, even if an ad hoc probe was written.
02 / THE REVIEW RECORD
What we actually inspected.
Source review has boundaries.
A clear record is more useful than a “safe” badge.
Material inspected
- plugins/claude-security/agents/patch-verifier.md (pinned complete source retained)
- plugins/claude-security/LICENSE
- Repository README and current official host configuration documentation
- plugins/claude-security/skills/claude-security/jobs/suggest-patches.md
- plugins/claude-security/agents/explore.md
- plugins/claude-security/LICENSE (overrides repository-wide licence for this plugin)
Our findings
- It rejects unstaged/untracked changes, derives the diff with external drivers disabled and checks the pre-change exploit before judging a fix.
- The behavior claim carries untested=true when no repository test exercises the changed code, even if an ad hoc probe was written.
- The exact-revision plugin licence is proprietary and expressly covers prompts, agent/skill definitions, workflows, code and documentation; the root Apache licence does not license this plugin for redistribution.
Not established by this review
- Source review only: no agent was installed or invoked. Task correctness, effective host restrictions and integration behaviour were not tested.
The review applies to the material and revision named here. A newer upstream release can change its behavior.
03 / PUT IT TO WORK
Use the role in your project.
- Read the plugin-specific proprietary licence at this revision. Obtain the plugin from Anthropic and use/modify it only within the applicable authorized Anthropic-product agreement; this catalogue provides no copy for installation.
- Read the pinned source and its licence, then configure an explicitly bounded host context.
- Run only inside the matching patch job with the patch base and companion explore/research components; review the project test command before granting execution.
- Supply a literal FINDING and the fix job’s validated scratch WORKSPACE.
Before you start
- Matching Claude Security plugin and orchestration
- Authorized repository and scratch-workspace context
- Project test dependencies already available
04 / FOLLOW THE EVIDENCE
The source trail.
Our notes are separate from the original resource.
Check upstream before adopting a new version.
https://github.com/anthropics/claude-plugins-official/blob/641d925db5a662e72f8650cb1b7670a59dab1abc/plugins/claude-security/agents/patch-verifier.md
Supports: Workflow, inputs, outputs and stated boundaries
- Applicable licence ↗Checked
https://github.com/anthropics/claude-plugins-official/blob/641d925db5a662e72f8650cb1b7670a59dab1abc/plugins/claude-security/LICENSE
Supports: Redistribution and attribution
https://code.claude.com/docs/en/sub-agents.md
Supports: Host loading and effective permission boundaries
- Pinned inspected source: plugins/claude-security/skills/claude-security/jobs/suggest-patches.md ↗Checked
https://github.com/anthropics/claude-plugins-official/blob/641d925db5a662e72f8650cb1b7670a59dab1abc/plugins/claude-security/skills/claude-security/jobs/suggest-patches.md
Supports: Dependency setup and source-review boundary
https://github.com/anthropics/claude-plugins-official/blob/641d925db5a662e72f8650cb1b7670a59dab1abc/plugins/claude-security/agents/explore.md
Supports: Dependency setup and source-review boundary