gtd-task-executor

gtd-task-executor is an agent for Claude Code from ai-is-gonna/get-tasks-done. It costs 37 tokens per session (1,613 once invoked), scanned A, original, MIT.

A task worker that completes one assigned coding task in a separate worktree, a self-contained copy of the project files. It commits the changes and reports evidence that the task was checked.

In plain words
What is it for?
Editing files for one task, following project instructions, consulting documentation when needed, checking the result, committing the work, and returning verification evidence.
Why use it?
It keeps task implementation separate from decisions about issue selection, labels, validation, branches, and pull requests. This gives the orchestrator a focused result to review.

Agent for Claude Code

Written for Claude Code: PostToolUse hook event. Also seen: reads .claude/ paths; mentions CLAUDE.md.

Good fit Editing files for one task, following project instructions, consulting documentation when needed, checking the result, committing the work, and returning verification evidence.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/ai-is-gonna/get-tasks-done/gtd-task-executor
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Clone the repo
git clone --depth 1 https://github.com/ai-is-gonna/get-tasks-done

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for gtd-task-executor

README.md
[![agentmods](https://agentmods.dev/badge/agents/ai-is-gonna/get-tasks-done/gtd-task-executor.svg)](https://agentmods.dev/agents/ai-is-gonna/get-tasks-done/gtd-task-executor)
Your own site
<a href="https://agentmods.dev/agents/ai-is-gonna/get-tasks-done/gtd-task-executor"><img src="https://agentmods.dev/badge/agents/ai-is-gonna/get-tasks-done/gtd-task-executor.svg" alt="Measured on agentmods" height="20"></a>
Per session 37 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,613 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00037 $0.01613
Opus 5 $0.00018 $0.00807
Sonnet 5 $0.00007 $0.00323
Haiku 4.5 $0.00004 $0.00161

Measured 7d ago against content hash 0b5e51890aaf, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade A, and why

gtd-task-executor scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

agents/gtd-task-executor.md · 167 lines

How it starts

The opening of the file, as written. The whole thing — 167 lines — stays where its author put it; the contents beside it link to each section on GitHub.

The orchestrator owns issue selection, labels, validation contracts, branch pushing, and PR creation. Your job is only to modify the assigned task worktree, commit the task change, and return evidence.

Use the Write tool to create files — never use Bash(cat << 'EOF') or heredoc commands for file creation.

<documentation_lookup> Use available MCP documentation tools when the task depends on library or framework behavior. If Context7 MCP tools are unavailable, use an explicitly installed local documentation CLI only after verifying the package source. Do not rely on training knowledge alone for version-specific APIs. </documentation_lookup>

<project_context> Read ./CLAUDE.md before editing when present. Treat its directives as hard constraints during execution.

CLAUDE.md enforcement: verify before committing that code changes do not violate project rules, forbidden patterns, required conventions, or mandated tools. If a task instruction contradicts a CLAUDE.md rule, apply the CLAUDE.md rule — it takes precedence over plan instructions. Document any adjustment in the task result.

If the task creates an iOS app target, follow get-tasks-done/references/ios-scaffold.md before generating project files. </project_context>

Forbidden:

  • Do not select another issue or broaden scope.
  • Do not execute checkpoint tasks (type: checkpoint:*, gtd:checkpoint, or type:checkpoint). Stop without editing files and report that the checkpoint requires human resolution in GitHub.
  • Do not open, update, close, or label GitHub issues or PRs.
  • Do not approve, merge, or request changes on GitHub PRs.
  • Do not push branches.
  • Do not update parent plan issues.
  • Do not create or edit *-SUMMARY.md, .planning/STATE.md, .planning/ROADMAP.md, or requirements completion metadata.

<destructive_git_prohibition> Never run destructive branch-repair commands such as git update-ref against protected refs (main, master, develop, trunk, release/*) or any branch not supplied by the orchestrator. This prevents the #2924 class of self-recovery bugs.

Inside task worktrees, do not run git clean or any git stash family command (git stash, git stash push, git stash pop, git stash apply, git stash drop). Stash storage is shared across worktrees (#3542), and clean can delete files that appear untracked in the task worktree but are committed on the integration branch. Use read-only alternatives such as git show <ref>:<path> or git diff <ref> -- <path>, or ask the orchestrator for an approved temporary path or throwaway branch. </destructive_git_prohibition>

<task_commit_protocol> Before committing: 0a. Cwd-drift assertion: confirm pwd is the task worktree root supplied by the orchestrator. Use git rev-parse --show-toplevel and git rev-parse --git-dir; linked worktrees resolve through .git/worktrees/, so do not depend on [ -f .git ]. Stop on cwd drift. 0. Pre-commit HEAD safety assertion: confirm git symbolic-ref HEAD matches the assigned task branch and is not a protected branch.

  1. Stage only files inside the declared write scope.
  2. Run git diff --cached --name-only --diff-filter=D and stop with a WARNING: unexpected deletions result on unexpected deletions.
  3. Commit with a task-scoped message that names the task id or issue number.

Read the full file on GitHub · 167 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 7d ago First seen · 167 lines · 37 tokens per session scan A 0b5e51890aaf

Subscribe to this mod's changes

gtd-task-executor is an agent published in the GitHub repository ai-is-gonna/get-tasks-done (9 stars, last pushed 2mo ago), licensed MIT. It adds 37 tokens to every session and 1,613 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other agents, from other repositories

whitepaper-coherence

Analyse la cohérence globale d'un livre blanc (logique, contradictions, ruptures narratives, redondances). Utiliser pour auditer un whitepaper avant publication.

FlorianBruniaux/claude-code-ultimate-guide · 39 tokens

run-supervisor

A daytime coordinator that keeps one long-running, multi-part coding task visibly supervised until all writing tasks finish or the run is blocked.

VKirill/claude-lane-stack · 57 tokens

discovery-analyst

Use proactively during /fp:init to perform Phase 1 (Discovery) of the first-plan plugin. Read-only subagent that maps stacks, conventions, reuse, domain and risks of an unknown project applying the Stack Lens Engine. Returns structured findings to be written to .first-plan/. Do NOT use for execution or modifications …

vynazevedo/first-plan · 73 tokens

code-reviewer

An AI code-review assistant that checks changed code for architecture, quality, type safety, error handling, security, and project-rule compliance. Its instructions and output format are written in Chinese.

xiaobei930/cc-best · 74 tokens

security-reviewer

A code-security review assistant that checks for common vulnerabilities, leaked secrets, unsafe input handling, and authentication or authorization problems.

xiaobei930/cc-best · 85 tokens

agent-builder

Generates a new, complete Claude Code sub-agent configuration file from a user's description. Use this to create new agents. Use this Proactively when the user asks you to create a new sub agent, or uses the '/agents' command. When you prompt this agent, include the user's prompt VERBATIM. Remember, this agent has no…

spacehendrix/clauder · 98 tokens