Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add Zintellix/Claude-Skills --skill swarmgit clone --depth 1 https://github.com/Zintellix/Claude-SkillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/zintellix/claude-skills/swarm)<a href="https://agentmods.dev/skills/zintellix/claude-skills/swarm"><img src="https://agentmods.dev/badge/skills/zintellix/claude-skills/swarm/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/zintellix/claude-skills/swarm"><img src="https://agentmods.dev/badge/skills/zintellix/claude-skills/swarm.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00138 | $0.01596 |
| Opus 5 | $0.00069 | $0.00798 |
| Sonnet 5 | $0.00028 | $0.00319 |
| Haiku 4.5 | $0.00014 | $0.00160 |
Grade A, and why
swarm scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 120 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Claude Swarm
Runs a parallel "swarm" of specialized subagents over a codebase, each scoped to a narrow category of issue, then synthesizes their raw findings into one deduplicated, severity-ranked action plan. Optionally executes fixes for approved findings.
Commands
/swarm audit [path]— run the full audit (default path: repo root)/swarm status— show the last audit summary from persisted state/swarm execute <severity|id>— fix findings matching a severity tier (critical/high/medium/low) or a specific finding id/swarm show <id>— show full detail for one finding
State file
All findings persist to .claude/swarm-audit.json at the repo root (create the
.claude/ dir if missing). This lets /swarm execute run in a later session without
re-auditing, and lets /swarm status answer instantly. Always read this file first
if it exists and the user references "the audit" or "findings" without re-running one.
Schema for this file is in references/output-schema.md.
/swarm audit [path] workflow
-
Scope check. If no path given, default to repo root, but if the repo is large (rough heuristic: >150 files or the user hasn't specified), ask the user to confirm scope or suggest a subdirectory. Don't silently audit a massive monorepo — cost and time scale with size.
-
Spawn the five audit agents in parallel. Use a single message with five Task tool calls so they genuinely run concurrently. For each, read the corresponding
agents/*.mdfile and pass its full contents as the subagent's instructions, plus:- the repo path/scope for this run
- the shared output schema from
references/output-schema.md - explicit tool restriction: read-only tools only (Read, Grep, Glob, and a
bash command runner if restricted to read-only commands like
git log,git blame, test runners in check-mode). Do NOT give audit agents Write/Edit — the audit phase must not touch files.
Agents to spawn:
agents/security.mdagents/performance.mdagents/tests.mdagents/architecture.mdagents/dead-code.md
What ships with it
9 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 120 lines · 138 tokens per session scan A a5d035750d92
swarm is a skill published in the GitHub repository Zintellix/Claude-Skills (4 stars, last pushed 10d ago), licensed MIT. It adds 138 tokens to every session and 1,596 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
architecture-review
Audits existing architecture for anti-patterns, scalability and reliability risks, and testability gaps. Graded findings with migration paths and a to-be diagram.
code-review
Structured engineering code review covering readability, complexity, test gaps, SOLID principles, and API consistency. Complements full-security-review with general code quality.
abd-review
ABD Tech Review agent: reviews code quality, architecture, naming, and error handling; writes severity-graded review artifacts.
review
Phase 5 of 5 — triage-first review orchestrator; dispatches up to 17 domain checks in parallel and compiles one report. Use only when the user asks to run Phase 5 or review a PR, branch, or set of changes — never trigger automatically after writing code.
architecture-review
Validates completeness and consistency of the project architecture against all GDDs. Builds a traceability matrix mapping every GDD technical requirement to ADRs, identifies coverage gaps, detects cross-ADR conflicts, verifies engine compatibility consistency across all decisions, and produces a PASS/CONCERNS/FAIL…
design-review
Reviews a game design document for completeness, internal consistency, implementability, and adherence to project design standards. Run this before handing a design document to programmers.