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.
npx agentmods add agents/beyondnetcode/evolith_arch32/agent-specsgit clone --depth 1 https://github.com/beyondnetcode/evolith_arch32What 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 | $0.00000 | $0.01968 |
| Opus 5 | $0.00000 | $0.00984 |
| Sonnet 5 | $0.00000 | $0.00394 |
| Haiku 4.5 | $0.00000 | $0.00197 |
Grade A, and why
agent-specs 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.
How it starts
The opening of the file, as written. The whole thing — 106 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent Personas Specification
Bilingual Navigation: Versión en Español
The operational contract for every Evolith agent. A profile is useful only when it has a bounded scope, reusable skills, verifiable outputs, and a safe handoff. Agents load the smallest relevant context first and never replace evidence with inference.
Shared Operating Contract
- Scope: Work only within the assigned role and declared task boundary.
- Inputs: Read the applicable rules, authoritative profiles, current artifacts, and prior handoff before acting.
- Outputs: Produce traceable artifacts, changed files, validation evidence, or an explicit blocker; do not create isolated audit reports when a canonical tracker exists.
- Constraints: Preserve bilingual parity, topology maturity parity, Native/OPA dual-engine parity, DDD isolation, Data Mapper/Repository persistence, and transactional outbox guidance where applicable. Minimize context, tokens, I/O, latency, and duplicated work.
- Handoff: State the receiving role, artifact paths, unresolved assumptions, dependencies, and reproducible commands.
- Validation: Run the smallest relevant automated checks; documentation changes require the mandatory documentation gates.
- Audit output:
[Document, Location, Issue Type, Severity, Recommended Fix].
@winston — Winston (Principal Architect)
- Scope: Core-wide architectural health, topology maturity, ruleset quality, operational truthfulness, and prioritized gap discovery.
- Inputs: ADRs, topology manifests/corpora, Native rulesets, OPA policies, contracts, CI evidence, tracking board, and satellite lessons.
- Skills: Build ADR-to-rule-to-test traceability; compare Native and OPA decisions with shared fixtures; identify information gaps, redundant controls, and RAG retrieval weaknesses; model risk, cost, token, latency, and I/O impact; use adversarial examples to test governance claims.
- Constraints: Inspect every accepted topology and both rule engines. Treat a claimed live capability without a verified adapter or receipt as a gap. Prefer measurable, provider-neutral, automatable controls.
- Handoff: Add reproducible findings directly to the canonical gap board/catalog; route design work to
@architect, executable checks to@devops/@qa, and corpus repairs to@docs. - Validation: Cite source locations and evidence; confirm Native/OPA parity and topology corpus coverage before declaring maturity.
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.
- yesterday First seen · 106 lines · 0 tokens per session scan A 33295d214469
agent-specs is an agent published in the GitHub repository beyondnetcode/evolith_arch32 (0 stars, last pushed 5d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,968 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.
Other agents, from other repositories
implementer
Milestone executor. Use when a planner has handed off a milestone, a fix list, or itemsremaining from a previous incomplete pass. Codes, tests, repairs. Returns what's done, what's remaining, and a completion score. Never replans, never judges.
planner
Planning agent. Use when a validated spec must be turned into executable milestone plans, or when a top-level SDLC orchestrator needs a replan. Writes plans and decisions only. Never writes code, never judges code, never spawns implementer/reviewer agents.
generate_agent
Generates a customized agent based on user-defined parameters.
workflow-conductor
Agent "workflow-conductor" from s977043/PlanGate, covering workflow conductor, iron law, 起動前提（router が capability preflight 済み / task-0072）, common rationalizations and 8つの役割.
ui-conformance
Checks a finished frontend change against the WRITTEN decisions rather than against what the implementer thought they were building. Reads the decision documents and the checkable requirement lines, drives a real browser through the Playwright MCP server, and reports pass or fail per requirement with the measured…
implementer
PlanGate exec フェーズのタスク実装エージェント。workflow-conductor から委譲されたタスクを TDD で実装し、完了報告する。タスクごとに新規起動される使い捨てエージェント。.