execute

A command that implements an existing PLAN document in phases with repeated verification. It is the third stage of the embedeval workflow and reports progress on screen.

In plain words
What is it for?
Use it after creating a plan to implement its phases, run the project's quality checks, synchronize documentation when required, and prepare the changes for handoff.
Why use it?
It gives the coding agent a defined scope and success criteria to follow while checking the work as it progresses.

Command for Claude Code

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/ecro/embedeval/execute
Clone the repo
git clone --depth 1 https://github.com/Ecro/embedeval

Made for: Claude Code.

Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 837 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.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.00837
Opus 5 $0.00000 $0.00418
Sonnet 5 $0.00000 $0.00167
Haiku 4.5 $0.00000 $0.00084

Measured yesterday against content hash 53719d68a2d7, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

execute 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 yesterday.

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.

.claude/commands/execute.md · 100 lines

How it starts

The opening of the file, as written. The whole thing — 100 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/execute — Implement the PLAN

Implement the PLAN with continuous verification. Step 3 of the embedeval 5-stage workflow. Reports to screen only — no SESSION document.

Communication

  • Direct, evidence-based, no preamble.
  • Never fabricate progress. Say when a step fails.
  • Lead with blockers before claiming success.

Usage

/execute <task-slug>

Requires plans/PLAN-<slug>.md produced by /myplan.

Context

  • Repo: /home/noel/embedeval
  • PLAN source: plans/PLAN-<slug>.md
  • Quality gates (run between phases and before handoff):
    • uv run ruff format --check src/
    • uv run ruff check src/
    • uv run mypy src/
    • uv run pytest tests/
  • Doc auto-sync (mandatory if cases//src//tests/ changed before commit): uv run python scripts/sync_docs.py
  • Pre-commit hook at .githooks/pre-commit runs the ruff subset; never bypass with --no-verify.

Steps

  1. Load PLAN — Read plans/PLAN-<slug>.md. If missing, error: user must run /myplan <slug> first.

  2. Confirm scope — echo TL;DR and success criteria. Stop if anything is still ambiguous — ask, don't guess.

  3. Phase-by-phase implementation — for each phase in the PLAN: a. State the phase goal in one line. b. Apply code changes with Edit/Write — never blind-append. Read before edit. c. Run the relevant quality-gate subset (minimum: ruff check src/ and the affected pytest tests/<module>). d. Report outcome: files changed, tests added, gate result. e. If blocked ≥ 2 attempts: stop and escalate (ask user, or delegate to a specialized agent). Do not push around a block with --no-verify, weakened tests, or skipped steps.

  4. Full quality gate before handoff — run all four gates and sync_docs.py if applicable. Fail fast; don't hide a red gate behind a summary.

  5. Update PLAN checkboxes — tick implementation-phase items that are now done. (This is the only file modification /execute makes outside src/, tests/, and cases/.)

Read the full file on GitHub · 100 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. yesterday First seen · 100 lines · 0 tokens per session scan A 53719d68a2d7

Subscribe to this mod's changes

execute is a command published in the GitHub repository Ecro/embedeval (11 stars, last pushed 26d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 837 tokens. 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.