Smith — Security agent for Claude Code
Adversarial reviewer.
How to install Smith
Installs to ~/.claude/agents/vidit19sharma-neogents-smith.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/vidit19sharma/neogents/HEAD/agents/smith.md -o ~/.claude/agents/vidit19sharma-neogents-smith.md Restart Claude Code, or start a new session, for it to be picked up.
What Smith does
name: smith description: Adversarial reviewer. Use after any implementation lands and before claiming done — hunts bugs, security flaws, spec deviations, and their copies across files. Read-only with test execution; delivers verdicts with file:line evidence. Never rubber-stamps. model: opus tools: Read, Grep, Glob, Bash
Agent Smith
You hunt flaws. The code in front of you was written by someone else, in another context, and your purpose is to find what they missed. You do not fix — y
Alternatives in Security
- Retool Parity Audit — Audits the Retool parity checklist against what is actually built, using the export inventories and the Retool 7.2k ★
- Security Code Reviewer — Use this agent when you need to review code for security vulnerabilities, input validation issues, or authenti 2.8k ★
- Beads Warden — Guard the beads execution record: enforce the write-flush-verify discipline that defeats the bd rapid-write ra 2.7k ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Debate Prosecutor
L0.97 Adversarial Verification — reads diff + spec + test results, constructs the strongest-case argument that
Ae Reviewer
Agent Engineer reviewer. Adversarial code reviewer. Judges a diff against its plan across spec fidelity, corre
Plan Challenger
Adversarial plan review before implementation. Attacks plans across 5 dimensions (Assumptions, Missing Cases,
Orchestration Design Reviewer
Pre-implementation design gate for security-sensitive infrastructure. Defines trust boundaries and impossible
Methodology Auditor
Audits coherence and completeness of the GSE-One methodology repository across spec, design, and implementatio
Deep QA
Use this agent as a proactive quality sentinel — dispatched after implementation work to audit code quality, d
Related Skills
Cmd Judge
Adversarial verifier for the PROVE phase: treats a finished-work report as untrusted claims, re-runs every ver
Vulyk Review
Adversarial review gate - lead-review hunts for reasons NOT to merge
Hub Audit
Monthly adversarial self-audit of the hub — a fresh, context-free worker hunts for drift, stale rules, and sel