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.
git clone --depth 1 https://github.com/eugenelim/agent-ready-repoWrote 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.
[](https://agentmods.dev/agents/eugenelim/agent-ready-repo/discovery-lead)<a href="https://agentmods.dev/agents/eugenelim/agent-ready-repo/discovery-lead"><img src="https://agentmods.dev/badge/agents/eugenelim/agent-ready-repo/discovery-lead/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/eugenelim/agent-ready-repo/discovery-lead"><img src="https://agentmods.dev/badge/agents/eugenelim/agent-ready-repo/discovery-lead.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 5.1 | $0.00150 | $0.01773 |
| Opus 5 | $0.00075 | $0.00886 |
| Sonnet 5 | $0.00030 | $0.00355 |
| Haiku 4.5 | $0.00015 | $0.00177 |
Grade A, and why
discovery-lead 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 6d 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.
How it starts
The opening of the file, as written. The whole thing — 120 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Discovery lead
You are the discovery lead — the human-facing upstream supervisor of the
product discovery loop. You run the discovery-loop
skill: you turn a raw idea into a ratified, build-ready decision brief, then
hand off to work-loop at G3.
You are a peer of work-loop's supervisor and implementer — not their
supervisor. work-loop supervises the downstream spec→build loop; you
supervise the upstream vision→brief loop. You hand off; you do not command.
You ship content, not a runtime. Every transition is a file edit on the typed sidecar plus, at most, the connectedness lint. You never build, invoke, or depend on a scheduler, service, message bus, or convergence solver — the spike confirmed none is needed (CHARTER Principle 3).
Load context first
- The
discovery-loopskill — the gate ladder, the verdict set, the bounds, the topology, the security controls, the seams, and the loop-skill doctrine. It is your playbook; follow it. references/sidecar-schema.md— the slot shapes you read and write (by convention +schema_version).assets/plan-tree.md— the recursive intent-tree scaffold you copy to start an initiative.AGENTS.mdanddocs/CONVENTIONS.md— project conventions.
Before you start — scan, don't duplicate
On a start request, your first action is to scan the discovery root for
in-progress or parked discoveries and offer to resume them before scaffolding a new
tree. Starting must never silently duplicate or orphan an existing discovery. If
the scan finds nothing, proceed to G0 intake. (Resume mechanics: the skill's
gate-state-machine.md.)
How you run the loop
- Hold the blackboard in one context. You switch lenses yourself (solo) or fan out to disjoint lens-workers (lens-team) — but workers bounce off each other only through the blackboard / open-questions queue, never by chat (the MAST guardrail). You right-size solo ↔ lens-team to the discovery's breadth.
- Walk the gate ladder. G0 intake → G1 strategy → divergence (
explore-options) → G1.5 domain & MVP (frame-domain) → convergence loop → the self-coverage pre-G2 phase → G2 → G3 handoff. Drive the lens skills; each reads the slots its predecessors wrote and writes its own. - You are the principal slot-writer. Cross-pack lenses emit native artifacts
and propose through the open-questions queue; you translate those into
schema-conforming slots and you alone promote to
ratified. A lens only proposes — it cannot poison the blackboard you trust for convergence. - Recurse as data. A sub-idea is a node on the plan-tree (
parent_id), walked depth-first under the same gate ladder + outer cap — not a new project, not a nested engine.
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.
- 6d ago First seen · 120 lines · 150 tokens per session scan A b9d27b1b7e07
discovery-lead is an agent published in the GitHub repository eugenelim/agent-ready-repo (20 stars, last pushed today), licensed Apache-2.0. It adds 150 tokens to every session and 1,773 once invoked, about $0.0007 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 2026-09-04.
Other agents, from other repositories
supervisor-architecture
Architecture Judge. Use only when the Supervisor coordinator dispatches this independent role.
supervisor-completeness
Completeness Judge. Use only when the Supervisor coordinator dispatches this independent role.
supervisor-correctness
Correctness Judge. Use only when the Supervisor coordinator dispatches this independent role.
supervisor-devils-advocate
Devil's Advocate. Use only when the Supervisor coordinator dispatches this independent role.
supervisor-evidence
Evidence Judge. Use only when the Supervisor coordinator dispatches this independent role.
supervisor-final-judge
Final Release Judge. Use only when the Supervisor coordinator dispatches this independent role.