Featured

Deploy OpenClaw in 60 seconds — 20% off logoDeploy OpenClaw in 60 seconds — 20% off

Launch OpenClaw on Hostinger in about 60 seconds and keep your agent live 24/7. Our referral link gives you 20% off, no coupon code needed.

Launch on Hostinger
Run your Hermes agent on Hostinger, fully managed logoRun your Hermes agent on Hostinger, fully managed

Launch Hermes on Hostinger in one click, fully managed, no VPS knowledge needed. Use code ZACAARON10 for 10% off.

Launch on Hostinger
Crawl and scrape any site into clean data, 10% off logoCrawl and scrape any site into clean data, 10% off

Firecrawl crawls and scrapes any site into clean markdown for your agent. Get 1,000 free credits, and new users get 10% off their first purchase.

Try Firecrawl free
Your own AI agent, running 24/7 with QwikClaw logoYour own AI agent, running 24/7 with QwikClaw

QwikClaw sets up and runs an always-on OpenClaw agent for you. One click, no config files, no server setup.

Deploy now
One API to scrape, enrich, and extract the internet. logoOne API to scrape, enrich, and extract the internet.

Context.dev gives your agents a single API to scrape, enrich, and extract live web data — no proxies, no parsers, no maintenance.

Start building free
SetupClaw: done-for-you OpenClaw for founders & exec teams logoSetupClaw: done-for-you OpenClaw for founders & exec teams

White-glove OpenClaw for founders and exec teams (4–50+ employees): we install, harden, integrate your tools, and maintain it — secured from day one.

Get it set up for you
SEO data APIs for your agent, $1 free credit logoSEO data APIs for your agent, $1 free credit

DataForSEO gives your agent live access to SERP results, keyword data, backlinks, and on-page SEO data through one API. New accounts get a $1 credit, good for up to 20,000 keyword or backlink lookups.

Try DataForSEO free
Reach 47,000+ AI builders

A flat monthly placement in front of developers actively installing AI tools. No lock-in, cancel anytime.

Advertise here
claude-issues logo

claude-issues

claude-issues

workflowClaude Codeby saifullahsaeed

Summary

Persistent markdown issue ledger with browser viewer, status audit, and cross-issue links for Claude Code projects.

Install to Claude Code

/plugin install claude-issues@claude-issues

Run in Claude Code. Add the marketplace first with /plugin marketplace add saifullahsaeed/claude-issues if you haven't already.

README.md

claude-issues

A persistent, markdown-based issue ledger Claude Code plugin — with a built-in browser viewer, status audit trail, and cross-issue links.

You spot issues during the day. Claude (in any future session) reads the ledger, checks for prior work touching the same files, picks issues, fixes them, and closes them. No database. Just markdown files in

.claude-issues/ inside your project, fully git-trackable.

Why

Claude Code sessions get cleared, restarted, or compacted. There is no shared memory of "what's already fixed in this project" vs. "what's still open." This plugin gives every project a durable, audit-friendly ledger that survives sessions, so Claude always knows the current state when you start a new conversation — and so you have a real history of how each problem was decided.

Install

In Claude Code:

/plugin marketplace add saifullahsaeed/claude-issues
/plugin install claude-issues@claude-issues

That's it. The CLI ships bundled inside the plugin — there's nothing to install from npm.

Quick start

In any project, just start using it — the ledger auto-creates on first use:

/claude-issues:issues add
/claude-issues:issues list
/claude-issues:issues view     # opens .claude-issues/_html/index.html in your browser
/claude-issues:issues serve    # for a clickable http://localhost URL

The bundled claude-issues skill auto-activates whenever you ask Claude to fix or investigate something: it reads the ledger first, checks for past similar fixes before starting new work, and posts a browser URL at the end of any reply that touched issues.

Depending on Claude Code version and conflicts, the slash command may also show as just /issues in autocomplete. Both forms work.

Statuses

| Status | Meaning | |---|---| | open | Active. Needs work. | | fixed | Verified fixed. Lives in .claude-issues/fixed/. | | wontfix | Closed without fixing (declined / out of scope). | | superseded | Replaced by a newer issue (superseded_by points forward). | | duplicate | Same as another issue (duplicate_of points to it). |

fixed, wontfix, superseded, and duplicate are all _closed_ states and contribute to the audit trail. Closed issues live in archive/ (except fixed, which keeps its own fixed/ folder for visibility).

Slash command

The /claude-issues:issues command in Claude Code wraps everything (may also be available as just /issues):

| Subcommand | What it does | |---|---| | add | Add an issue (interactive); warns if similar past issues exist | | list / list fixed / list archive / list all | List issues | | show <id> | Print an issue's full markdown | | fix <id> "<note>" | Mark fixed; move to fixed/ | | wontfix <id> "<note>" | Close without fixing | | reopen <id> | Move any closed issue back to open/ | | note <id> "<text>" | Append a timestamped progress note | | link <id> supersedes <old> | New issue replaces an older one | | link <id> duplicate-of <other> | Mark <id> as a duplicate | | link <id> related <other> | Add a bidirectional related link | | view [id] | Open the browser viewer (file://) | | serve [port] | Start a local HTTP server for a clickable http://localhost URL | | init | Optional — every other command auto-creates the ledger if missing |

IDs accept 1, 001, or ISSUE-001 interchangeably.

Browser viewer

Every CLI write also regenerates a static HTML site at

.claude-issues/_html/index.html. Each command prints a clickable

file://… URL at the end of its output:

View: file:///abs/path/.claude-issues/_html/ISSUE-008.html

Click it (or run /issues view) to browse a styled, navigable ledger with severity / status badges and links between superseded, duplicate, and related issues. The skill instructs Claude to always post this URL at the end of any reply that touches issues.

On-disk layout

.claude-issues/
  INDEX.md                       # auto-generated entry point for Claude
  _html/                         # browser viewer (gitignored)
    index.html
    ISSUE-001.html …
  open/
    ISSUE-003-slow-dashboard-query.md
  fixed/
    ISSUE-001-login-broken-on-safari.md
  archive/                       # wontfix · superseded · duplicate
    ISSUE-002-…-superseded.md
    ISSUE-005-…-duplicate.md
    ISSUE-006-…-wontfix.md

Each issue file has YAML frontmatter (id, title, status, severity,

files, created, fixed_at, closed_at, supersedes,

superseded_by, duplicate_of, related) and a markdown body with description, repro steps, and an auto-appended "Fix notes" section.

How Claude uses it

The bundled claude-issues skill instructs Claude to:

1. Read .claude-issues/INDEX.md first to learn project state. 2. Before fixing, scan the ledger (open + fixed + archive) for prior work touching the same files or keywords. If it finds a match, ask the user whether the new request is a regression (→ reopen), an intentional different approach (→ supersede), or a duplicate (→ link). 3. Read the full issue file, including any prior fix notes. 4. Record progress with note <id> "..." as it works. 5. Verify the fix, then close with the right command:

  • fix for verified fixes
  • wontfix for declined / out-of-scope
  • link --duplicate-of if it turns out to be a duplicate
  • add --supersedes <old> if it's a different fix for the same problem

6. Post the View: file://… URL at the end of every reply.

Contributing / building locally

git clone https://github.com/saifullahsaeed/claude-issues
cd claude-issues
npm install
npm run build      # rebuilds plugins/claude-issues/bin/cli.cjs

The bundled CLI at plugins/claude-issues/bin/cli.cjs is committed so plugin users don't need to build anything.

License

MIT

Related plugins

Browse all →