Skip to main content
skill CX Codex

ds-ledger — Claude Code skill: what it does, how to install

This Claude Code skill logs decisions during long or unattended work. It is triggered for multi-phase tasks, work that a person will review later, or with /ds-ledger.

MIT 545 installs ★ 320 updated 2026-10-28

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add aicodeking/deepseek-stack --skill ds-ledger -a claude-code
Open on GitHub

What ds-ledger does

The skill creates a TSV ledger, records the goal and completion condition in its first row, then adds a row after each decision with the phase, choice, reason, evidence, and outcome. Before each step, the agent rereads the goal and the last ten rows to check whether the next action still serves it.

Evidence must point to something that can be opened or rerun, such as a commit hash, file:line, or artifact path. Each cell stays on one line, and old rows are left intact; corrections are appended. Reverts and failed work are logged too. Three consecutive rows without progress toward the goal stop the run, which is then reported. At the end, the agent adds a summary row and points to the ledger file.

What it helps with

  • track decisions during a long agent run
  • check each next step against the original goal
  • record why work was reverted or dropped
  • leave evidence for a later human review

Example prompts for the agent

  • Track decisions for this multi-phase task with /ds-ledger.
  • Run autonomously and keep evidence for each decision using /ds-ledger.
  • Create a decision log for a person to review later: /ds-ledger.

How to install ds-ledger

1Skills CLI (Vercel)

npx skills add aicodeking/deepseek-stack --skill ds-ledger -a claude-code

2Manually: put the files in place

~/.claude/skills/ds-ledger/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/ds-ledger/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about ds-ledger

How do I install ds-ledger in Claude Code?

Run npx skills add aicodeking/deepseek-stack --skill ds-ledger -a claude-code. The skill file is at ~/.claude/skills/ds-ledger/SKILL.md.

What does ds-ledger do?

It keeps a TSV record of decisions and checks the stated goal before each step. The log includes reasons, evidence, outcomes, reversions, and failed work.

When does ds-ledger stop a run?

It stops and reports when three consecutive rows show no progress toward the goal.

Source text

SKILL.md · MIT
---
name: ds-ledger
description: "Keep a one-row-per-decision log during long or unattended runs, and re-read it before each step so the run does not drift from its goal. Use for multi-phase or autonomous work, work a human will review later, or /ds-ledger."
---

# ds-ledger

A long run forgets its own reasons. A short written ledger keeps the goal in view and lets a reviewer trust the result.

## Steps

1. Create one TSV file, for example `.ds-ledger.tsv`, with the header `ts	phase	decision	why	evidence	result`. Put the goal and done condition in the first row.
2. Before each step, re-read the goal row and the last ten rows. Check that the next step still serves the goal. If it does not, log the drift and correct course.
3. After each real decision, append one row:
   - ts: ISO 8601 time.
   - phase: the phase or workstream.
   - decision: what you chose or did, one line.
   - why: the reason in plain words.
   - evidence: a pointer, never prose. A commit hash, `file:line`, a command and its result, or an artifact path.
   - result: the outcome, such as "tests pass", "reverted", "open".
4. Keep each cell on one line with no tabs.
5. Log reverts and dropped work too, with the reason. Failures are the most useful rows.
6. At the end, add a summary row and point the user to the file. Commit it only if a reviewer needs it.

## Stop when

- The run ends and the summary row is written.
- If three rows in a row show no progress toward the goal, stop the run and report.

## Rules

- Every row needs evidence you can open or rerun. "Looked fine" is not evidence.
- Write rows the way you would tell a teammate, plain and concrete.
- Do not rewrite old rows. Append corrections.

© aicodeking · MIT · aicodeking/deepseek-stack

Similar skills

All skills for Agent setup: memory, context, orchestration →

From the same set29

The whole set →
ds-design ds-design — a Claude Code skill for setting a distinct visual direction for frontend work. It runs when creating a page, app screen, or component, and when restyling an existing interface. CX Codex +1 548installs ds-review-loop ds-review-loop — a skill for Claude Code that runs an author and a reviewer in an iterative loop until they agree on the code changes. CX Codex +1 548installs ds-unslop This Claude Code skill rewrites prose and code comments in plain, specific English. It activates while editing replies, docs, READMEs, PR descriptions or commit messages, and on /ds-unslop. CX Codex +1 547installs ds-blueprint Defines types, signatures, and module boundaries before writing code bodies. It triggers on /ds-blueprint or when asked to design the structure first. CX Codex +1 546installs ds-explain Skill explains code logic and architectural reasons in plain language for Claude Code, Codex, and Cursor. Runs during subsystem onboarding or via /ds-explain. CX Codex +1 545installs ds-lessons ds-lessons is a Claude Code skill that turns recurring agent mistakes in a repository into hard checks, from types to linters. It runs after user corrections, repeated bugs, or long sessions. CX Codex +1 544installs ds-sharpen The ds-sharpen Claude Code skill refines your prompts, deepens tasks, and stress-tests plans for AI agents. It runs before starting vague or high-stakes work to identify potential weaknesses. CX Codex +1 542installs ds-tdd The ds-tdd Claude Code skill guides bug fixes through a failing test, then a minimal code change. It activates for TDD or regression-test requests and bugs with a simple local test target. CX Codex +1 542installs

Project sponsors

Companies that keep this analytics open