agentsclimarketplace

Handoff

Skill dayfinggg/claude-code-codex-skills/claude/skills/handoff

Drop-in Agent Skills, subagents, output styles, hooks & config templates for Claude Code (~/.claude) and OpenAI Codex CLI (~/.codex). Copy, restart, done.

Install
npx -y skills add dayfinggg/claude-code-codex-skills --skill handoff

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • 18 days oldThe repository was created 18 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Create a compact, secret-safe continuation brief for another Claude Code session or engineer. Use only when the user explicitly asks to hand off, compact, transfer, pause and resume, or continue substantial work elsewhere. Do not use for normal final reports, short tasks, or automatic status updates.

SKILL.md

4.3 KB, 773 tokens by cl100k_base, as published. Nobody here has run it

Create Handoff

Preserve the live, resumable core of the work. Reference durable sources of truth instead of copying them.

Read references/evidence.md when deciding what state must be externalized for a fresh session. Do not delegate handoff authorship: one owner must reconcile the current workspace, decisions, and verification into a single source of continuation truth.

Respect the Handoff Contract

  • Create a handoff only on explicit request.
  • Do not continue implementation, make unrelated changes, or reopen settled decisions while preparing it.
  • Save the brief in the operating system's temporary directory by default so it does not become repository maintenance burden.
  • Save it in the workspace only when the user specifies a path or the repository has an established authorized handoff location.
  • Return the brief directly in the response instead of creating a file when the user requests response-only output.

Reconstruct the Current Truth

Verify the current state from available evidence rather than relying only on conversation memory:

  • active instructions, plan, specification, and accepted decisions;
  • workspace and version-control state;
  • changed and untracked files;
  • relevant command, test, build, and validation results;
  • open tool processes or runtime state;
  • external task, issue, PR, document, and deployment references.

Distinguish completed work, partially completed work, proposed work, and unverified claims.

Preserve What the Next Task Needs

Capture:

  • the objective and requested outcome;
  • current scope and explicit exclusions;
  • decisions already made and the reasons that constrain future work;
  • the current implementation state and material files;
  • completed verification with exact results;
  • failures, uncertainties, blockers, and user input still required;
  • the next concrete action;
  • commands needed to resume safely;
  • relevant paths, symbols, URLs, commits, branches, task IDs, and artifacts;
  • ownership of pre-existing or unrelated workspace changes.

Include assumptions only when they remain active and material.

Reference, Do Not Duplicate

Link to existing plans, specifications, ADRs, issues, diffs, commits, logs, and documentation. Summarize only the live thread that those artifacts do not already contain.

Do not copy:

  • settled detail available in a cited source;
  • long command output or conversation history;
  • obsolete attempts with no effect on the next action;
  • generic project information a fresh session can discover normally.

Preserve rationale for consequential decisions so the next session does not relitigate them.

Protect Sensitive Information

Remove credentials, tokens, keys, cookies, private personal data, internal secrets, and sensitive command output. Replace necessary references with safe descriptions of where authorized access is expected.

Do not include secret values even if they appeared earlier in the conversation or terminal.

Structure the Brief

Use only applicable sections:

# Handoff

## Objective
## Current state
## Decisions and rationale
## Changes
## Verification
## Open issues
## Next action
## References

Keep the brief concise enough for a fresh session to read fully before acting. Use absolute local paths and direct external URLs where appropriate.

Write and Verify

When creating a file:

  • use a unique descriptive Markdown filename;
  • write it to the resolved temporary or user-specified location;
  • reread it to confirm completeness and path accuracy;
  • scan it for obvious secret patterns and remove sensitive content;
  • report the final path and a one-sentence prompt the user can give the next session.

Do not claim the handoff is complete if the workspace state or verification results could not be inspected; state the missing evidence.

Completion Standard

Finish only when a capable fresh session can identify what is being done, why current decisions were made, what is already complete, what remains, and the exact next action without reading the full prior conversation.

Gives 0 of the 12 instructions most agent orchestration skills give in 773 tokens

Counted across 742 of the 995 authors here whose files we hold, read 2026-08-06

  • run the full test suite after integrating changesin 53 of 742, across 20 files
  • reference existing artifacts by path or URLin 52 of 742, across 22 files
  • dispatch one agent per independent problem domainin 50 of 742, across 17 files
  • verify fixes do not conflictin 45 of 742, across 13 files
  • include a suggested skills section in the documentin 45 of 742, across 15 files
  • redact sensitive informationin 41 of 742, across 11 files
  • save to the temporary directory of the operating systemin 39 of 742, across 9 files
  • tailor the document to user-provided focus argumentsin 39 of 742, across 9 files
  • spot check agent changes for systematic errorsin 34 of 742, across 7 files
  • write a handoff document summarising the current conversationin 31 of 742, across 6 files
  • assign each agent a specific scopein 23 of 742, across 8 files
  • provide specific scope and clear goalin 23 of 742, across 5 files

Said here and by no other author read

  • create handoff only on explicit request
  • stop implementation while preparing handoff
  • distinguish completed, partial, and unverified work
  • capture the objective and current scope
  • link to durable sources instead of copying them
  • remove credentials, tokens, keys, and secrets

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.