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 lm203688/swarmlabs-engine-kit --skill vv-gategit clone --depth 1 https://github.com/lm203688/swarmlabs-engine-kitWrote 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/lm203688/swarmlabs-engine-kit/vv-gate)<a href="https://agentmods.dev/skills/lm203688/swarmlabs-engine-kit/vv-gate"><img src="https://agentmods.dev/badge/skills/lm203688/swarmlabs-engine-kit/vv-gate/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/lm203688/swarmlabs-engine-kit/vv-gate"><img src="https://agentmods.dev/badge/skills/lm203688/swarmlabs-engine-kit/vv-gate.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 | $0.00107 | $0.02679 |
| Opus 5.5 | $0.00043 | $0.01072 |
| Sonnet 5.5 | $0.00021 | $0.00536 |
| Haiku 4.5 | $0.00011 | $0.00268 |
Grade A, and why
vv-gate scanned grade A with 1 finding 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 27d 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
default `Python-urllib/3.x`.** The response body is `error code: 1010`. How it starts
The opening of the file, as written. The whole thing — 204 lines — stays where its author put it; the contents beside it link to each section on GitHub.
V&V Gate — independent adjudication for numeric claims
Most "verification" in agent pipelines is the agent grading its own homework. This skill is a read-only client to a gate you do not control, so the answer can actually be "no".
Core distinction, and the reason this skill exists at all: reproducible ≠ valid. A fully reproducible experiment can be fully wrong. SwarmLabs does not sell reproducibility (git already does that). It sells adjudication against held-out ground truth plus literature anchors.
Two different questions — do not mix them
| Question | Command | Where the truth lives |
|---|---|---|
| What is the current trust state of scenario X? | check X |
vendor's published gate ledger (static) |
| Are my predictions on this held-out set correct? | verify X --pred p.json |
held-out set whose y the vendor holds, you don't |
check answers "is this scenario's surrogate trustworthy at all".
verify answers "you are wrong". Only the second one is a gate on your agent.
Prerequisite
Zero dependencies. Python 3.8+ stdlib only. No API key, no numpy, no engine clone. The script is self-contained:
python skills/vv-gate/scripts/vv_gate.py selftest
selftest is a live smoke test (it hits the public endpoints, it is not a
unit test). Expected shape:
[1] benchmark index: total=62 points=3110 policy_src=engine/verification_gate.py::thresholds()
[2] gate ledger: total=62 blocked=0
[3] verify(<first-key>) self-oracle: HTTP 200 verdict=PASS r2=1 gate=PROCEED
[4] anchors: 6 entries
SELFTEST OK
If it prints SELFTEST FAILED, stop and read references/gate-semantics.md
before wiring it into anything.
Exit codes are the product
Verification that cannot block is not verification, so the exit code is the contract:
| Code | gate | Meaning |
|---|---|---|
0 |
PROCEED |
claim cleared |
3 |
PROCEED_WITH_HUMAN_CHECK |
needs a human signature |
2 |
BLOCK_AUTONOMOUS_ACTION |
must halt the autonomous action |
4 |
— | unreachable — edge block / network failure; no verdict exists |
1 |
— | local call/parameter error (bad JSON, unfilled template, non-JSON reply) |
What ships with it
4 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.
- 27d ago First seen · 204 lines · 107 tokens per session scan A e60d4124147f
vv-gate is a skill published in the GitHub repository lm203688/swarmlabs-engine-kit (0 stars, last pushed 29d ago), licensed MIT. It adds 107 tokens to every session and 2,679 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-18.
Other skills, from other repositories
llamaguard
Meta's 7-8B specialized moderation model for LLM input/output filtering. 6 safety categories - violence/hate, sexual content, weapons, substances, self-harm, criminal planning. 94-95% accuracy. Deploy with vLLM, HuggingFace, Sagemaker. Integrates with NeMo Guardrails.
master-buddhaghosa
A reference-based guide to Buddhaghosa, a fifth-century Theravāda Buddhist scholar, and the meditation and doctrine explained in his works. It covers the Three Trainings—ethics, concentration, and wisdom—along with meditation subjects and stages of insight.
bluesky-engagement
Automated Bluesky reply monitoring and draft queueing for ThumbGate's acquisition engagement loop. Polls the AT Protocol notifications endpoint every 15 minutes, writes human-reviewable draft replies into a queue file, and never auto-posts without sign-off. Trigger when the user asks about Bluesky replies, engagement…
agent-architect-kit
Status: shipped 2026-04-21. Test-backed. CLI-automatable. Safe to recommend publicly.
Agent Memory
Give your agent persistent cross-session memory — recall what went wrong on similar tasks before starting, then capture whether the work succeeded or failed so repeated mistakes auto-promote into prevention rules. Runs locally via the ThumbGate MCP server (recall, capturefeedback, preventionrules, feedbackstats) — no…
context-engineering-checklist
HARD checklist before new skills, MCP tools, hooks/gates, plugins, or multi-agent workflows. Maps Hugging Face Context Course units 1–5 onto ThumbGate. Auto-invoke when creating SKILL.md, PreToolUse hooks, gate patterns, MCP tools, Grok workflows, slash commands, or when user says context engineering, HF context…