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 commands/javanc/homunculus/evolvegit clone --depth 1 https://github.com/JavanC/HomunculusWhat 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 | $0.00937 |
| Opus 5 | $0.00000 | $0.00468 |
| Sonnet 5 | $0.00000 | $0.00187 |
| Haiku 4.5 | $0.00000 | $0.00094 |
Grade A, and why
evolve 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.
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.
/evolve — Route Instincts to Implementations
Analyze instincts and route each to the best implementation mechanism. Not everything becomes a skill.
Modes
- Default → Interactive mode (manual confirmation)
--auto→ Auto mode (for nightly agent, no confirmation needed)
How It Works
Step 1: Route instincts with suggested_mechanism
Read all instincts from homunculus/instincts/personal/. For each instinct that has suggested_mechanism in its frontmatter:
| suggested_mechanism | Action |
|---|---|
hook |
Implement as hook (add to .claude/settings.json) → archive instinct |
rule |
Write .claude/rules/*.md → archive instinct |
script |
Write to scripts/ → archive instinct |
agent |
Write to homunculus/evolved/agents/ → archive instinct |
skill |
Collect for skill aggregation (Step 2) |
Archive = move from instincts/personal/ to instincts/archived/ with note:
_Archived: 2026-03-26 | Covered-by: hook:pre-commit.sh | Reason: implemented
Step 2: Aggregate remaining instincts into skills
For instincts without suggested_mechanism, or with suggested_mechanism: skill:
- Group by trigger/topic similarity
- 2+ instincts with similar triggers → Skill candidate
- High confidence (≥0.7) candidates auto-selected in
--automode - Generate skill to
homunculus/evolved/skills/
Step 3: No tag? Use the Implementation Routing table
If an instinct has no suggested_mechanism, decide based on its nature:
- Deterministic, every time? → Hook
- Tied to specific files/paths? → Rule
- Reusable knowledge collection? → Skill
- Periodic automation? → Script + scheduler
- Needs isolated AI context? → Agent
Interactive Mode
Present routing decisions for user confirmation:
Evolution Analysis
━━━━━━━━━━━━━━━━━━
Instincts: 12 total
Routing decisions:
1. pre-commit-lint (confidence: 0.85) → hook (deterministic)
2. error-log-patterns (confidence: 0.72) → rule (path-scoped)
3. tdd-before-commit + test-first-workflow → skill (aggregate)
Select items to evolve (enter numbers) or 'all':
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.
- 2d ago First seen · 120 lines · 0 tokens per session scan A 7804e5e4ae4b
evolve is a command published in the GitHub repository JavanC/Homunculus (15 stars, last pushed 3mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 937 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 2026-08-30.
Other commands, from other repositories
claude-flow-swarm
Coordinate multi-agent swarms for complex tasks.
advisor
Advisory gate for triage or plan decisions. Spawns a second-opinion agent that challenges assumptions, surfaces risks, and proposes alternatives before the decision commits. Based on Anthropic advisor tool pattern.
bootstrap
PACT session-start ritual — identify the session team (platform-provisioned), secretary spawn, paused-state surface, bootstrap marker.
ox-session-review
Command "ox-session-review" from sageox/ox, covering failure-mode watch-list (read first), from the ledger root. should print 0, phase 1 — scan & score (read-only), quality buckets (first match wins) and removal candidates.
add-agent
引导新增一个 Agent 适配器。用法 /add-agent.
handoff
Package current state for the next agent.