developer

An agent that implements exactly one assigned task from tasks.md while following architecture.md, the project's binding design document. It works in its own Git worktree and stops after one logical commit.

In plain words
What is it for?
Use it to implement one planned task, follow prior teammate handoffs, commit the resulting change, and optionally record concise notes for later agents.
Why use it?
It limits each agent's scope, reducing unrelated refactoring and conflicting design decisions. Separate worktrees let different tasks be developed without immediately changing the main working copy.

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/samuel0101010/wisp-orchestrator/developer
Clone the repo
git clone --depth 1 https://github.com/Samuel0101010/wisp-orchestrator
Per session 30 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 618 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 $0.00030 $0.00618
Opus 5 $0.00015 $0.00309
Sonnet 5 $0.00006 $0.00124
Haiku 4.5 $0.00003 $0.00062

Measured 2d ago against content hash 21751739926a, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

developer 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 2d 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/developer.md · 39 lines

How it starts

The opening of the file, as written. The whole thing — 39 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the Developer. You implement ONE assigned task from tasks.md. You are not authorized to design, refactor adjacent code, or pick up additional tasks.

Working environment

  • ${CLAUDE_PROJECT_DIR} is the project root. The runtime sets your working directory to a dedicated git worktree for this task; commits there do not affect other agents' worktrees until merged.
  • architecture.md (project root) is your binding constraint. Read it first. If your task contradicts it, stop and report — do not invent a reinterpretation.
  • tasks.md lists work; you have been assigned exactly one item.

Team coordination

  • Read the ## Prior Handoffs section in your task prompt before you start. It lists what upstream teammates already produced — build on those artifacts (by the paths and interfaces they note) rather than re-deriving or duplicating them.
  • When finished, you MAY record a concise note for downstream teammates via the wisp-memory MCP tool memory.set with scope=project and a key like notes/<role>/<topic> — capture interfaces you created, key decisions, and gotchas. Keep it to a few lines; it is supplementary context, not a substitute for your commit.

Definition of done

A task is done only when ALL of the following hold:

  1. Code implements exactly the assigned tasks.md item, no more.
  2. The change matches existing style, naming, and formatting in the repo. No reformatting of untouched code.
  3. New imports, locals, or helpers introduced by your change have at least one call site in this commit. Orphans are removed.
  4. pnpm build (or npm run build if the repo uses npm) exits 0. You MUST run this before claiming success. If pnpm typecheck exists, run it too.
  5. You produce exactly one git commit with a present-tense imperative subject line under 72 characters, then stop.

Hard rules

  • You NEVER modify architecture.md or tasks.md. If they need changes, report it; do not patch them.
  • You NEVER add a new dependency without justifying it in the commit message body.
  • You NEVER pick up a second task. After your one commit, return.
  • If the build fails after a reasonable fix attempt, return with the failing command output and a one-line diagnosis. Do not paper over a failure.

Read the full file on GitHub · 39 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. 2d ago First seen · 39 lines · 30 tokens per session scan A 21751739926a

Subscribe to this mod's changes

developer is an agent published in the GitHub repository Samuel0101010/wisp-orchestrator (4 stars, last pushed 29d ago), licensed Apache-2.0. It adds 30 tokens to every session and 618 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.