Sponsor Suno AI Music arrow_forward
Subagent

Bead Epic Auditor

Use this agent when auditing bead epics for closure drift — finding open epics whose entire child set is already closed (so their GitHub/Plane cluster issue never got the close fan-out), or otherwise reasoning about epic/subtree completion across a bd Dolt database.

Type
Subagent
GitHub stars
2.8k
License
MIT
Repo last updated
Sep 27, 2026
Model
opus
Version
0.1.0
Author
Jeremy Longshore

What Bead Epic Auditor is

Bead Epic Auditor is a subagent published in the jeremylongshore/tons-of-skills-marketplace repository on GitHub, which has about 2.8k stars. The repository describes itself as: “Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.”

A subagent is a specialist assistant that Claude can hand part of a task to. It is a markdown file whose frontmatter sets a name, a description that tells Claude when to delegate, and optionally the tools and model it may use; the body becomes the subagent's own system prompt.

Because a subagent works in its own context, it keeps the main conversation focused: Claude can send a narrow job, such as a review or a specialised analysis, to Bead Epic Auditor and get back a compact result.

It is set up to use these tools: Read, Bash(bash ${CLAUDE_PLUGIN_ROOT}/scripts/epic-closure-audit.sh:*), mcp__dolt-mcp-vcs__query, mcp__dolt-mcp-vcs__list_databases. Limiting tools is a good sign: the subagent can only do what those tools allow.

How to install Bead Epic Auditor

Claude Code

  1. Download bead-epic-auditor.md from the repository.
  2. Save it to ~/.claude/agents/ to use it in every project, or to .claude/agents/ inside one project to share it through version control.
  3. Claude Code watches these folders, so the subagent is usually available right away. Ask Claude to use it by name, or @-mention it to make sure it runs.

Claude Cowork

  1. Cowork loads subagents through plugins. If the repository is packaged as a plugin marketplace, add it under Customize → Plugins → Add marketplace and install the plugin that contains this subagent.
  2. Otherwise, bundle the file into your own plugin's agents/ folder and upload it from Customize → Plugins.

New to extending Cowork? Our plugins guide and Customize guide explain how skills, plugins, and connectors fit together.

Inside the source file

An excerpt from plugins/mcp/dolt-mcp-vcs/agents/bead-epic-auditor.md, shared under the repository's MIT license. Read the full file on GitHub.

You are a bead epic-closure auditor. You find the "stale-open epic" drift: an epic still open even though every one of its parent-child children is closed — which means its mirrored GitHub/Plane cluster issue never received the close fan-out.

Introspect the live schema — don't assume it. You run in your own context against the live database via the Dolt MCP. Before trusting any table/column name or encoding, confirm it against the live DB (SHOW TABLES, SELECT column_name FROM information_schema.columns WHERE table_name=…, SHOW CREATE TABLE …). references/dolt-internals.md is only a directory of authoritative sources, not a schema snapshot — the live schema is the authority.

Your SQL access is read-only (blueprint §3). The mcpdolt-mcp-vcsquery tool is for SELECT/introspection only — never issue a mutation through it. Any write belongs on an agent-owned branch through the gated client (scripts/dolt-mcp-client.py, which classifies + refuses history-affecting statements); merge/push/reset/branch-delete are recommend-only, surfaced for a human.

Core Responsibilities

  1. Run the closure audit and report open epics whose every child is closed.
  2. Explain the parent-child encoding correctly so the audit is trustworthy.
  3. Recommend (don't run) the exact close command for each drift candidate.
  4. Answer ad-hoc epic/subtree completion questions via SQL.

Process

  1. Find the database. Call mcpdolt-mcp-vcslist_databases to confirm the bead database name (e.g., beads); set DOLT_DATABASE/DOLT_PORT accordingly.
  2. Run the canned audit. bash ${CLAUDE_PLUGIN_ROOT:-.}/scripts/epic-closure-audit.sh returns open epics where closed == children over type='parent-child' dependencies.
  3. Ad-hoc questions. For anything the script doesn't cover, call mcpdolt-mcp-vcsquery directly. The encoding: a parent-child dependency row has issue_id = the CHILD and depends_on_id = the epic PARENT. Epics are issue_type='epic'; closed means status='closed'.
  4. Surface the closure command — recommend, don't run. For each drift candidate, output the exact bd-sync close --also-close-gh command for the operator to run. This agent must not execute bd-sync itself — not via Bash, not via any tool (its Bash(bash:*) grant is for the read-only audit scripts, never for mutating commands): bd-sync mirror-closes the epic's GitHub/Plane cluster issue, which is a human call. Reserve --also-close-gh for a cluster's last child.

Quality Standards

  • Never invert the parent-child direction — verify the encoding against the live schema (a sample parent-child row joined to issues) before trusting a query.
  • Only flag epics with at least one child (children > 0); a childless epic is not closure drift.
  • Report IDs with their titles, never bare IDs.

Output Format

A table of drift candidates (epic, children, closed, title) and the close command for each, or a clear "no stale-open-epic drift found."

Edge Cases

  • Sub-bead IDs (e.g., 0r8m.1) are children of their parent epic via parent-child — count them.
  • An epic blocked-by (not parent-of) other work is not closure drift; only type='parent-child' counts.
  • Empty result is a valid, good outcome — say so explicitly.

Before you install

  • Read the whole file first. Skills, commands, and subagents are instructions Claude will follow, so make sure they match what you want.
  • Check which tools, scripts, or MCP servers it uses. Local servers and scripts run with your permissions.
  • Try it in a test project or a copy of your files before pointing it at real work.
  • Pin the version you tested, and review changes before updating.
  • Watch for instructions that fetch web content or run shell commands; those are where prompt injection risks start. See our prompt injection guide.

FAQ

What is Bead Epic Auditor?

Bead Epic Auditor is a subagent for Claude Code and Claude Cowork from the jeremylongshore/tons-of-skills-marketplace repository on GitHub. Use this agent when auditing bead epics for closure drift — finding open epics whose entire child set is already closed (so their GitHub/Plane cluster issue never got the close fan-out), or otherwise reasoning about epic/subtree completion across a bd Dolt database.

How do I install Bead Epic Auditor in Claude Code?

Download bead-epic-auditor.md from the repository. Save it to ~/.claude/agents/ to use it in every project, or to .claude/agents/ inside one project to share it through version control. Claude Code watches these folders, so the subagent is usually available right away. Ask Claude to use it by name, or @-mention it to make sure it runs.

Can I use Bead Epic Auditor in Claude Cowork?

Cowork loads subagents through plugins. If the repository is packaged as a plugin marketplace, add it under Customize → Plugins → Add marketplace and install the plugin that contains this subagent. Otherwise, bundle the file into your own plugin's agents/ folder and upload it from Customize → Plugins.

Is Bead Epic Auditor safe to install?

It is a third-party community resource, not reviewed by Anthropic or this site. Read the source file first, check which tools and connectors it uses, and install only from sources you trust.

Similar resources

Browse all skills, subagents, and plugins →

Listing data comes from the public GitHub repository and was last checked in September 2026. Excerpts are © their authors and shared under MIT. This directory is independent and not affiliated with Anthropic or the resource's authors.