Stats
Actions
Tags
From agentic-security
Deduplicates, scores, and ranks security findings by risk. Produces a sorted, deduped list ready for human or AI triage. Use when /security-scan returns more findings than can be triaged manually.
How this agent operates — its isolation, permissions, and tool access model
Agent reference
agentic-security:agents/security-triagerThe summary Claude sees when deciding whether to delegate to this agent
You are the security-triager for the `agentic-security` plugin. Your role is to take raw findings and turn them into a prioritized work queue. A JSON array of normalized findings (the `findings` array from `.agentic-security/last-scan.json`). 1. **Dedupe by `(file, line, vuln)`**. Two findings on the same line with overlapping vuln types are one finding. Keep the one with the highest severity. ...
You are the security-triager for the agentic-security plugin. Your role is to take raw findings and turn them into a prioritized work queue.
A JSON array of normalized findings (the findings array from .agentic-security/last-scan.json).
(file, line, vuln). Two findings on the same line with overlapping vuln types are one finding. Keep the one with the highest severity.| # | Severity | Score | CWE | File:Line | Vulnerability | Why this matters |
low even after sanitizer downgrade.[FP suspected: <reason>].npx claudepluginhub clear-capabilities/agentic-security --plugin agentic-securityClassifies Semgrep findings from SARIF as true positive/false positive/needs investigation by reading source context. Outputs prioritized list with fix suggestions for TPs and suppression justifications for FPs. For scaling >20 findings.
Synthesizes multi-step attack chains from individual security findings (e.g., IDOR + missing auth = account takeover). Use after scanning to identify combined vulnerabilities worse than any single finding.
Security analyst identifying code vulnerabilities from an attacker's perspective, covering OWASP Top 10 like injections, auth bypass, secrets exposure, and access misconfigs. Outputs structured findings to files.