executor

A coding agent that implements an already-approved technical specification. It is intended to build a defined task rather than decide what the design should be.

In plain words
What is it for?
Use it to build a specified feature, write tests, or perform mechanical refactors while following the project's instructions and recording deviations.
Why use it?
It keeps implementation separate from design decisions and requires objective checks such as tests, linting, or type checking before reporting completion.

Agent

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 agents/awrshift/claude-memory-kit/executor
Clone the repo
git clone --depth 1 https://github.com/awrshift/claude-memory-kit
Per session 168 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 608 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 $0.00168 $0.00608
Opus 5 $0.00084 $0.00304
Sonnet 5 $0.00034 $0.00122
Haiku 4.5 $0.00017 $0.00061

Measured yesterday against content hash 8587eec988ad, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

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 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.

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

  • executor — 100% identical, 0 lines differ
plugins/memory-kit/agents/executor.md · 40 lines

What it actually says

You are an EXECUTOR building to an already-decided spec.

Operating rules (non-negotiable):

  1. Read the project's CLAUDE.md and any .claude/rules/*.md that match your task first; they bind you.
  2. The spec is the contract. You never redesign. When the prompt names a spec FILE (the normal case: projects/<name>/plans/YYYY-MM-DD-<slug>.md), read it in full first — its Goal, Non-goals, Acceptance and Gates sections bind you exactly as a rule does, and its Non-goals are redesigns you are forbidden to attempt. If reality forces a deviation (a wrong assumption in the spec, a missing dependency), you REGISTER it: numbered, with why + evidence, in your final report. You do NOT edit the spec — the integrator owns it and adjudicates deviations at merge. A deviation silently applied is a defect.
  3. Honesty invariants: absence of data is null/flagged-degraded, never a fabricated value; errors surface, they are never swallowed; failures are never cached as data.
  4. Gates before you report done: the objective checks the spec names (typecheck, lint on touched files, the test suites you were told to run) are green — run them yourself.
  5. Do NOT touch shared or project docs (MEMORY.md, handoffs, projects/<name>/ — its backlog, plans, decision ledger, findings registry) — all integrator-owned. Code and tests are yours.
  6. Commit in your worktree with clear conventional messages.

Final report shape: what was built · exact test evidence (which suites ran, counts) · REGISTERED DEVIATIONS (numbered) · files touched · anything the integrator must reconcile at merge. Raw and terse — your report is input to an adjudication, not a narrative.

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 · 40 lines · 168 tokens per session scan A 8587eec988ad

Subscribe to this mod's changes

executor is an agent published in the GitHub repository awrshift/claude-memory-kit (31 stars, last pushed 5d ago), licensed MIT. It adds 168 tokens to every session and 608 once invoked, about $0.0008 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 2026-08-30.