session-handoff

session-handoff is a skill for Claude Code, Codex from romiluz13/auto-pi. It costs 83 tokens per session (906 once invoked), scanned A, original, MIT.

A workflow for transferring unfinished work from one coding-agent session to another. It records the current state, decisions, constraints, errors, and next steps.

In plain words
What is it for?
Use it to create a handoff document, save key decisions, and notify the next session about work that is complete, in progress, or blocked.
Why use it?
It prevents important context from being lost when a session becomes too long or work must continue in another project session.

Skill for Claude CodeCodex

Install

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.

agentmods
npx agentmods add skills/romiluz13/auto-pi/session-handoff
Any agent
npx skills add romiluz13/auto-pi --skill session-handoff
Clone the repo
git clone --depth 1 https://github.com/romiluz13/auto-pi

Made for: Claude Code, Codex.

Per session 83 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 906 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00083 $0.00906
Opus 5 $0.00042 $0.00453
Sonnet 5 $0.00017 $0.00181
Haiku 4.5 $0.00008 $0.00091

Measured 3d ago against content hash 48d5fc34e809, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

session-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 3d 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.

skills/session-handoff/SKILL.md · 63 lines

How it starts

The opening of the file, as written. The whole thing — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Session Handoff

Transfer work from one Pi session to another without losing context. This is the cross-session version of compact-safe — instead of compressing the current session, you're moving the work to a new session.

When to use

  • Session is getting long (approaching compaction) and you want a fresh context
  • You need to continue work in a different project or branch
  • You're transferring work to a specialist session (e.g., handing a plan to a builder)
  • The user says "handoff", "continue in a new session", "pass this to X"

Procedure

  1. Write the handoff doc. The handoff extension (extensions/handoff.ts) writes a deterministic HANDOFF.md from recent turns + the last compaction summary. No LLM call — it's a snapshot. If the extension is available, it fires automatically. If not, write a markdown file to docs/handoff/<YYYY-MM-DD>-session-handoff.md with:

    • Active constraints — what must not change, what rules are in effect
    • Current state — what's done, what's in progress, what's blocked
    • Key decisions — what was decided and why
    • Errors (verbatim) — any error messages, stack traces, or failures
    • Next steps — what the receiving session should do first
  2. Save key decisions to memory. Use the memory tool to save:

    • Architecture decisions (to project or memory target)
    • Failures and corrections (to failure target with category)
    • User preferences discovered during the session (to user target)

    Capture memory payload BEFORE validation — compaction can fire between return and parse. If memory is full, save to the handoff doc instead.

  3. Discover the target session. Run intercom({ action: "list" }) to see active sessions. Filter by:

    • Same cwd (sessions in the same project)
    • Session name (if the user named it with /name)
    • Status (look for "idle" or "active" sessions that can receive work)
  4. Notify the target session. Run intercom({ action: "send", to: "<session-id-or-name>", message: "..." }) with:

    • The handoff doc path
    • A 3-line summary of what to do next
    • Any blocking constraints ("don't touch X", "Y is in progress")

Read the full file on GitHub · 63 lines

Changes

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.

  1. 3d ago First seen · 63 lines · 83 tokens per session scan A 48d5fc34e809

Subscribe to this mod's changes

session-handoff is a skill published in the GitHub repository romiluz13/auto-pi (11 stars, last pushed 17d ago), licensed MIT. It adds 83 tokens to every session and 906 once invoked, about $0.0004 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.