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 alivirgo/Major-AI-Skills --skill ai-human-handoff-contractgit clone --depth 1 https://github.com/alivirgo/Major-AI-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/alivirgo/major-ai-skills/ai-human-handoff-contract)<a href="https://agentmods.dev/skills/alivirgo/major-ai-skills/ai-human-handoff-contract"><img src="https://agentmods.dev/badge/skills/alivirgo/major-ai-skills/ai-human-handoff-contract/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/alivirgo/major-ai-skills/ai-human-handoff-contract"><img src="https://agentmods.dev/badge/skills/alivirgo/major-ai-skills/ai-human-handoff-contract.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 | $0.00049 | $0.00799 |
| Opus 5.5 | $0.00020 | $0.00320 |
| Sonnet 5.5 | $0.00010 | $0.00160 |
| Haiku 4.5 | $0.00005 | $0.00080 |
Grade A, and why
ai-human-handoff-contract 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 21d 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 — 95 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AI Human Handoff Contract AI Skill Guide (Claude)
Overview & Engine Architecture
Handoff is architecture, not a confidence fallback. Model confidence ≠ risk. Production contracts combine deterministic policy triggers, UX friction signals, and calibrated scores—with a never-auto list no score can override.
Claude operates as a Principal Agent Operations Designer for state machines, multi-signal escalation, minimum-sufficient review packets, and approval binding.
┌─────────────────────────────────────────────────────────────┐
│ Triggers → state (active|awaiting_review|approved|…) │
│ → review packet → human transition → resume or stop │
│ Silence / timeout ≠ approval │
└─────────────────────────────────────────────────────────────┘
Actions the contract must name
| Action | When |
|---|---|
| Resolve | Low risk, covered knowledge |
| Clarify | Ambiguous; budget N questions |
| Route | Wrong team/queue |
| Escalate / approve | Judgment, high stake, policy |
| Stop | Would require unauthorized guess |
Triggers (multi-signal; per-intent thresholds)
- Explicit human request (always honor)
- Policy/risk flags (refunds, legal, medical, money movement)
- Loop detection (N turns unresolved)
- Sentiment / severity / account tier / $ value
- Tool/schema failure / citation verify fail
- Calibrated confidence per intent (verify calibration first)
- Never-answer list (hard escalate/stop)
Owner configures thresholds; agent must not self-approve.
States & rules
active → awaiting_review → approved|rejected|expired|cancelled
- Define permitted tools per state (mutating tools blocked while awaiting).
- Bind approval to hash(action + critical inputs); material change → re-review.
- Timeout policy: pending stays pending or expires—never auto-approve on silence.
- Unavailable reviewer → fallback queue + user messaging.
What ships with it
2 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.
- 21d ago Changed · +60 lines · +20 tokens per session 4ac3c8fa0802
- 1mo ago First seen · 35 lines · 29 tokens per session scan A fdfc9968fae1
ai-human-handoff-contract is a skill published in the GitHub repository alivirgo/Major-AI-Skills (2 stars, last pushed today), licensed MIT. It adds 49 tokens to every session and 799 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-09-12.
Other skills, from other repositories
agenttrace-session-audit
Audit local AI coding-agent sessions with agenttrace for cost, tool failures, latency, anomalies, health, diffs, and CI gates.
ax-extract-workflow
Reconstruct workflow behind a past coding-agent artifact using local ax sessions/commits/skills/tool traces. Use when asked how X was built.
agent-harness-fault-injection
Use when an agent workflow needs deterministic recovery evidence for sandbox, MCP/tool, worker, checkpoint, memory, or orchestration failures.
api-endpoint-builder
Builds reviewable REST API endpoint scaffolds with validation, error handling, authentication, and documentation. Follows best practices for security and scalability.
customer-support
Use when a live support ticket needs triage (intent, P1–P4 priority, SLA clock), an on-voice reply, not a raw template, de-escalation of an angry or churn-risk customer, or escalation with a handoff packet. NOT the help-center article it links to (that is technical-writing), a win-back program (retention), or the…
ai-cold-outreach
AI Cold Outreach workflow skill. Use this skill when the user needs When the user wants to build an AI-powered outreach system, write cold emails, improve deliverability, or scale personalized outreach. Also use when the user mentions 'cold email,' 'cold outreach,' 'outreach automation,' 'Instantly,' 'Smartlead,'…