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 3xachris/3xa-harness --skill handoffgit clone --depth 1 https://github.com/3xachris/3xa-harnessWrote 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/3xachris/3xa-harness/handoff)<a href="https://agentmods.dev/skills/3xachris/3xa-harness/handoff"><img src="https://agentmods.dev/badge/skills/3xachris/3xa-harness/handoff/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/3xachris/3xa-harness/handoff"><img src="https://agentmods.dev/badge/skills/3xachris/3xa-harness/handoff.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00119 | $0.01272 |
| Opus 5 | $0.00060 | $0.00636 |
| Sonnet 5 | $0.00024 | $0.00254 |
| Haiku 4.5 | $0.00012 | $0.00127 |
Grade A, and why
handoff 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 9d 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 — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Handoff
A handoff is read by someone who was not in the room — not a colleague filling gaps from memory of the conversation, a reader with none. decision-log is written for the cold read of one decision at a time, appended the moment it happens. A handoff is written for the cold read of the whole state at once: a single wide snapshot, taken the moment work changes hands, of what the log's entry-by-entry trail doesn't by itself add up to — everything true right now, in a shape built to survive the writer's own memory being gone at the other end.
1. The coverage checklist
Four categories, each answered or marked none — leaving one out reads identically to having checked it and found nothing, so only a written none proves the second:
- Running work and its reasoning. Every thread currently open, its status word from the fixed set (
WORKING,GATE-HOLD,STOPPED,DONE,CLOSED-FAILED— defined inhonest-closeout) and the one-line reason it sits there. Where a thread is a frozenworkorder, point at the order instead of redescribing it. - Already-spent assets. Anything paid for, provisioned, or installed that outlives this session — how to start it, and any red line attached to using it: a quota, a licence limit, an action it must never take.
- Cross-thread dependencies. Which open items block or feed which others, so the next reader doesn't start the wrong one first.
- Unresolved debt. Known problems nobody has closed, each with enough of a trail that the next reader can pick it up rather than rediscover that it exists.
2. Write for the reader, not the writer
The most common way a handoff fails: written as a log of what got done this session, in the order it happened. A narrative like that has no edge that stops it short of complete — it just stops wherever the writer's memory of the session runs out, and what's missing is exactly the part the writer wasn't touching when they sat down to write: someone else's open thread, an asset from three sessions back, a debt nobody's mentioned lately. Write against the four categories instead, and write each fact as itself rather than as something someone did — a fact stated plainly reads the same whether the cold reader arrives in an hour or in a month.
What ships with it
1 file 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.
- 9d ago First seen · 74 lines · 119 tokens per session scan A 9a8cc971ad41
handoff is a skill published in the GitHub repository 3xachris/3xa-harness (17 stars, last pushed 9d ago), licensed MIT. It adds 119 tokens to every session and 1,272 once invoked, about $0.0006 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 2026-08-30.
Other skills, from other repositories
auto-resume
Recover active change and next action from artifacts. Use on fresh session with existing work.
agent-carnet
Use this skill when the user asks to save, recall, find, or organize notes. Triggers on: 'remember this', 'save this', 'note this', 'what did we discuss about...', 'check the notebook', 'find in carnet'. Also use proactively when discovering findings worth preserving across sessions.
remember
Review reusable project knowledge and decide what belongs in project memory, notepad, or durable docs.
mnemos
Task-scoped memory lifecycle — typed MnemoGraph prevents lossy context compaction by treating facts/decisions/code-refs/handoffs as distinct node types with per-type eviction policies.
grimoire
Use when the user says 'update context', 'update claude', 'save library', or after significant project changes.
ln-31-performance-optimizer
Profiles and improves a measured latency, throughput, CPU, memory, or I/O problem. Not for speculative tuning or cosmetic refactoring.