Claude Code plugin that scans your repo or a diff with a team of agents, verifies each finding, and drafts patch files.
Foe: Breach Spider, a security hole an attacker could use.
Foes
Party · Context
The Breach Spider blocks the way.
What really happensIt stands for a security hole an attacker could use.
Rogue: Context 100/100.
The spider strikes through a gap nobody guarded. Its bite lingers.
What really happensAn attacker gets in through a hole nobody checked, and the harm spreads until it is patched.
Rogue takes 11. Rogue: Context 89/100. Rogue: Poison.
The hero sounds the Claude Security horn inside the session.
What really happensRun /claude-security and pick scan codebase, scan changes, or suggest patches. The scan states its time and token cost first.
Breach Spider takes 16.
Scouts map the keep, then hunters spread through every wing.
What really happensAgents map the architecture, build a threat model, and hunt across each component. Vendored code such as node_modules is skipped.
Mapper: Context 40/40. Threat modeler: Context 40/40. Hunter: Context 40/40. Hunter: Context 40/40.
The Breach Spider lashes at Mapper.
What really happensThe subagent takes the hit in its own context. The main context stays clean.
Mapper takes 11. Mapper: Context 29/40. Mapper: Poison.
The summons work the problem, each in its own context.
What really happensEach subagent does its own model and tool work, apart from the main conversation.
Breach Spider takes 7. Breach Spider takes 7. Breach Spider takes 7. Breach Spider takes 5.
The summons return with short reports.
What really happensOnly the subagent's final summary comes back to the main agent.
Rogue takes 2. Rogue takes 2. Rogue takes 2. Rogue takes 2. Rogue: Context 81/100. Rogue: Insight. Mapper returns. Threat modeler returns. Hunter returns. Hunter returns.
Each claimed hole must survive a second band sent to disprove it.
What really happensIndependent verifiers try to prove every candidate a false positive. Only findings with a confirmed exploit path reach the report.
Verifier: Context 40/40. Verifier: Context 40/40.
The summons work the problem, each in its own context.
What really happensEach subagent does its own model and tool work, apart from the main conversation.
Breach Spider takes 7. Breach Spider takes 6.
The summons return with short reports.
What really happensOnly the subagent's final summary comes back to the main agent.
Rogue takes 2. Rogue takes 2. Rogue: Context 77/100. Verifier returns. Verifier returns.
Smiths forge a patch for each hole, away from your camp.
What really happensSuggest patches writes one F<n>.patch per finding in a scratch copy of the repo, reviewed by other agents. Nothing is applied for you.
Breach Spider takes 19.
The spider's lair is charted, with a fix ready for each web.
What really happensYou get CLAUDE-SECURITY-RESULTS.md plus JSONL and SARIF files. Apply a patch with git apply when you choose.
Breach Spider takes 26. Breach Spider falls.
The Breach Spider falls.
Claude Code
/plugin install claude-security@claude-plugins-official
/reload-pluginsVersion 0.12.0. Proprietary Anthropic license: internal use with Anthropic products only. Run /claude-security to scan the codebase, scan changes, or suggest patches. A scan asks you to confirm its time and token cost. Results go to CLAUDE-SECURITY-<timestamp>/ as Markdown, JSONL and SARIF. Patches are never applied for you. Needs Python 3.9+. Meant for code you control; for repos you do not trust, the README suggests running inside sandbox-runtime. Sends usage counts unless telemetry is off.
Quest Find and fix exploitable holes in a web app you own before it ships
Claude Opus 5.5
Anthropic Security Guidance Plugin
Semgrep Guardian
Anthropic Claude Security Plugin
StackHawk Agent SkillsThis spell has no recorded sessions in the curator's logs. Tried it? Send a spell report.