What is execute-tasks?

A command for carrying out an approved software work plan from start to finish. It follows the listed tasks, checks the result against the testing plan, and reviews the implementation.

It helps resume a documented task, implement its requirements and design, run verification, and record or continue the work using the project's task log.

It removes the need to manually coordinate implementation, testing, and review for each work item.

execute-tasks — Command for coding agents by MadaraUchiha-314: A command for carrying out an approved software work…

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.

agentmods

npx agentmods add commands/madarauchiha-314/the-loop/execute-tasks

Clone the repo

git clone --depth 1 https://github.com/MadaraUchiha-314/the-loop

Or install the-loop, the plugin that ships this one along with the rest of its 2 skills, 37 commands, 2 hooks.

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.

README.md

[![agentmods](https://agentmods.dev/badge/commands/madarauchiha-314/the-loop/execute-tasks.svg)](https://agentmods.dev/commands/madarauchiha-314/the-loop/execute-tasks)

Your own site

<a href="https://agentmods.dev/commands/madarauchiha-314/the-loop/execute-tasks"><img src="https://agentmods.dev/badge/commands/madarauchiha-314/the-loop/execute-tasks.svg" alt="Measured on agentmods" height="20"></a>

Per session 31 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 1,446 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. Scan, not verified.

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.

Model	Per session	Once invoked
Fable 5	$0.00031	$0.01446
Opus 5	$0.00015	$0.00723
Sonnet 5	$0.00006	$0.00289
Haiku 4.5	$0.00003	$0.00145

Measured 4d ago against content hash f3f9fedb207e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

execute-tasks 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 4d 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.

commands/execute-tasks.md · 93 lines

How it starts

The opening of the file, as written. The whole thing — 93 lines — stays where its author put it; the contents beside it link to each section on GitHub.

the-loop: execute-tasks `$ARGUMENTS`

Implement a work item by executing its approved tasks.md against requirements.md, design.md and testing-plan.md — the implementation → verification → review portion of the loop. A slice of /the-loop:work-on; work-on remains the superset.

Read the the-loop skill, reference/workflow.md, reference/context.md and reference/tooling.md first. Load .the-loop/harness-config.yaml; read every custom instruction doc it registers (customInstructions.docs, in order — the operator's conventions and styles, reference/instructions.md, and the-loop instructions to confirm each one resolved); apply any per-task overrides from the spec front-matter.

Start clean. Entering implementation crosses a phase boundary: apply contextManagement.phaseBoundary (default clear) so execution runs against the locked spec files read from disk, not the drafting conversation (plan-mode style).

Steps

Locate & load the spec. Resolve $ARGUMENTS to docs/specs/<id>/ and read requirements.md, design.md, testing-plan.md, tasks.md, and execution-log.md. Use the log's phase and tasks' checkmarks to resume rather than restart.
Implementation (implementation). Execute the task DAG autonomously in dependency order. Tick each task in tasks.md (- [ ] → - [x]) as it completes. Maintain execution-log.md: append progress and run tests (unit/integration per config) at the task checkpoints — self-check as you go. Use the configured tooling; same commands as CI. Keep the ticket phase label in sync. After each completed task, manage the context window: checkpoint first (checkmark, log entry with a concrete Next, WIP committed/noted), then reset per contextManagement.taskBoundary (default compact). Mid-task, compact only — never clear. Never reset without the checkpoint. See reference/context.md.

Stay monitorable. Register the session for webhook/poll routing (the-loop sessions register, see the skill's reference/automation.md). When the ticketing provider is not GitHub (Jira, …), the PR is the monitorable GitHub object: once the PR exists, add the routing.autoExecuteLabel (from the operator's CLI config, decision-032; default the-loop: auto-execute) to the PR directly and register against the PR's own ref (github:OWNER/REPO#<pr-number>), so PR activity still resumes this session. Every PR you open is recorded against the work item in the same step as opening it — the-loop sessions link-pr --work-item github:OWNER/REPO#N --pull-request <pr-number> — because a PR the-loop authored carries none of the linkages the router can infer, and without the binding its review comments reach no session at all. If the item takes more than one PR, label every one of them, record every one of them, and list them all in the execution log's Pull requests table — one PR merging does not end the work item.

Read the full file on GitHub · 93 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.

4d ago First seen · 93 lines · 31 tokens per session scan A f3f9fedb207e

Subscribe to this mod's changes

execute-tasks is a command published in the GitHub repository MadaraUchiha-314/the-loop (4 stars, last pushed today), licensed MIT. It adds 31 tokens to every session and 1,446 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.

Code review Planning Testing agentic-coding claude-code claude-plugin claude-skills cursor loop

execute-tasks

Getting it into your agent

Wrote this? Show the measurements

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

the-loop: execute-tasks `$ARGUMENTS`

Steps

What this file has done since we first saw it

Other commands, from other repositories

gaai-audit

brainstorm

execute-plan

gaai-ask

gaai-status

gaai-update

execute-tasks

Getting it into your agent

Wrote this? Show the measurements

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

the-loop: execute-tasks $ARGUMENTS

Steps

What this file has done since we first saw it

Other commands, from other repositories

gaai-audit

brainstorm

execute-plan

gaai-ask

gaai-status

gaai-update

the-loop: execute-tasks `$ARGUMENTS`