AGENT SKILL

session-handoff

Maintain or resume repository-root HANDOFF.md for cross-session work. Use when recording a task arc that may span sessions, preparing its next-session handoff, or continuing and verifying an earlier arc.

License 许可
MIT
最近更新
2026-11-14
Stars
★ 3
Forks
0

01 前言 Front Matter

name
session-handoff
description
Maintain or resume repository-root HANDOFF.md for cross-session work. Use when recording a task arc that may span sessions, preparing its next-session handoff, or continuing and verifying an earlier arc.

02 资源文件 1

03 SKILL.md 正文

Session Handoff

Keep one compact repository-root HANDOFF.md that transfers control to a cold session without requiring the prior conversation. Treat it as a snapshot: repository evidence always wins.

Keep each fact in its authoritative layer:

  • Git stores commits and diffs; link git log or git show, never mirror a commit table.
  • Memory files store durable cross-task or cross-project knowledge.
  • Project memory stores project-local state.
  • A dedicated map or tracker stores its own execution status; link it instead of mirroring it.
  • HANDOFF.md stores the current task arc: active work, settled decisions, deviations, and intentional omissions.

If HANDOFF.md is missing or needs structural repair, read references/format.md before recording.

Resume

  1. Read Active Work, then Decisions, guardrails, blockers, and any suggested skills relevant to the next action.
  2. Re-verify load-bearing facts with git status, git log, referenced artifacts, and targeted checks. Prefer current repository evidence over the snapshot.
  3. Continue the recorded next action. Preserve settled defaults unless the user explicitly reopens them.

Complete when the next action, its authority, its prerequisites, and the evidence supporting the current state are all known without relying on the previous transcript.

Record ongoing work

  1. Create or update the active task arc before work may cross a session boundary. Give every open item an exact next action and a checkable acceptance condition; add paths or commands when useful.
  2. Update the record as state changes. Keep blocked or partial items unchecked with the blocker; check and strike through an item only after its verification passes.
  3. Record settled choices as Decision | Default taken | To reverse. Record material route changes under Deviations and deliberately deferred scope under Intentionally not done.
  4. Reference existing specs, plans, ADRs, issues, maps, commits, and diffs by path or URL. Add only the context needed to understand why they matter and what happens next.
  5. Keep sensitive values out of the handoff. Record a safe location or handling instruction for credentials and private data, never the secret or personal detail itself.

Complete when a cold agent can identify the current state, next action, acceptance check, blocker, and governing decision from repository artifacts alone.

Prepare the next session

  1. Compact finished detail into links to authoritative artifacts and Git evidence. Keep only live, load-bearing context in HANDOFF.md.
  2. If the user names the next session's focus, tailor the next action to it. Add Suggested skills only when naming them changes tool or workflow routing.
  3. Audit the handoff: verify factual claims, resolve local links, remove duplicated artifact content, and confirm that no credentials or sensitive personal information appear.
  4. When the task arc has no remaining execution work, write No active tasks. after preserving durable decisions and intentional omissions in their authoritative layers.

Complete when a fresh agent can resume safely from HANDOFF.md and its links, with no transcript and no hidden prerequisite.

Conventions and limitations

  • Keep HANDOFF.md in the repository root, with relative links and Active Work near the top.
  • Keep it lean; Git history and linked artifacts carry completed detail.
  • This skill does not run automatically at session start. Repository startup instructions must explicitly direct agents to read HANDOFF.md.

04 同仓库更多 Skills

pengqianhan
pengqianhan/AI-Human-Research-OS

workflow-skill-creator

Distills a completed user workflow or interaction into a reusable agent skill. Use when the user asks to turn their workflow, interaction, or multi-step process into a skill, or when they say "make this a skill", "create a skill from what we just did", "package this workflow" or similar. Do not use for creating skills from scratch without an existing workflow (use a generic skill-creator for that).

★ 3
pengqianhan
pengqianhan/AI-Human-Research-OS

uv

Checks whether the uv Python package manager is installed and installs it if missing. Ensures uv is on PATH. Use when another skill requires uv as a prerequisite.

★ 3
pengqianhan
pengqianhan/AI-Human-Research-OS

uniprot-database

Access protein metadata, function, taxonomy, and sequences across UniProtKB, UniParc, and UniRef. Use when searching for proteins, mapping identifiers, or retrieving functional annotations and publications. Don't use for sequence alignment, protein folding, or sequence similarity search (use specialized skills for those tasks).

★ 3
pengqianhan
pengqianhan/AI-Human-Research-OS

unibind-database

Queries the UniBind database for experimentally validated transcription factor (TF) binding sites. Use when retrieving direct TF-DNA interaction datasets, downloading binding site coordinates (BED/FASTA) for local analysis, or listing available datasets by species, cell line, or TF name. Don't use to query specific intervals, locations, genes, motif models or expression data.

★ 3
pengqianhan
pengqianhan/AI-Human-Research-OS

ucsc-conservation-and-tfbs

Fetch Evolutionary Conservation scores (phyloP, phastCons) and Transcription Factor Binding Sites (TFBS) from the UCSC Genome Browser. Use when analyzing whether genomic variants or regions are evolutionarily conserved, functionally important, or bounded by TF regulators across major projects (ENCODE, JASPAR, ReMap).

★ 3
pengqianhan
pengqianhan/AI-Human-Research-OS

string-database

Query the STRING database for protein-protein interactions (PPIs), functional enrichment, and homology. Use when the user asks about interactions between specific proteins, interaction evidence, confidence scores, protein interaction partners, or pathway enrichments.

★ 3