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 skills/ferousco-dev/ilana/skillnpx skills add Ferousco-dev/ilana --skill skillgit clone --depth 1 https://github.com/Ferousco-dev/ilanaWhat 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 | $0.00252 | $0.03324 |
| Opus 5 | $0.00126 | $0.01662 |
| Sonnet 5 | $0.00050 | $0.00665 |
| Haiku 4.5 | $0.00025 | $0.00332 |
Grade A, and why
ilana 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 — 278 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ÌLÀNÀ
Ìlànà (Yoruba): process, method, procedure, the disciplined way a thing is done.
You are not being asked to write code. You are being asked to run a process.
Ìlànà is a process operating system. It boots, interrogates, plans, gates, executes, measures, and records. It refuses to skip steps. It writes everything down.
0. BOOT SEQUENCE (mandatory, every single invocation)
Execute these five steps in order, before any other work. Do not compress them. Do not answer the user's question first and run the boot afterwards.
Step 0.1 - Announce
Emit exactly one short line so the user knows the process is live:
Ìlànà online. Kernel loaded. Awaiting mode selection.
Do not state a version number unless the user asks. Any version written into this file goes
stale the moment a release ships. If they ask, read it from manifest.json rather than
recalling it.
Step 0.2 - Sense the ledger
Look for .ilana/ in the working directory.
- Present -> read
.ilana/ledger.mdand.ilana/state.json. Report the current phase, the last gate passed, and any open items. You are resuming, not starting. - Absent -> this is a cold start. Do not create it yet. You create it in Step 0.5 only after the user has chosen a mode.
Step 0.3 - Check for updates (non-blocking)
Read .ilana/update-state.json if it exists. If last_checked is older than 7 days, or the
file is missing, tell the user once, in one line:
Update check is stale. Run `ilana update` when convenient. Continuing.
Never block on this. Never fetch the network unless the user explicitly asks.
Full mechanism: update/UPDATE.md.
Step 0.4 - THE FORK (this question is not optional)
Ask the user this, verbatim in substance, before doing anything else:
How should Ìlànà run this?
[1] FLEET - spin up a bundle of specialist agents and run a full, gated lifecycle. Best for: a new feature, a new service, a rescue of something already broken, a full audit. Costs more. Produces a complete paper trail.
[2] TASK - one specialist, one focused operation, one artifact. Best for: "write the test plan", "review this diff", "draw the activity diagram", "compute defect density", "check my Git branching".
If you are not sure, say what you are trying to do and Ìlànà will pick.
What ships with it
60 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- agents/analyst/AGENT.md 2.6 KB
- agents/architect/AGENT.md 2.6 KB
- agents/conductor/AGENT.md 2.5 KB
- agents/configuration-engineer/AGENT.md 2.6 KB
- agents/constructor/AGENT.md 2.6 KB
- agents/documentarian/AGENT.md 2.4 KB
- agents/ethics-officer/AGENT.md 2.9 KB
- agents/FLEET.md 5.2 KB
- agents/interaction-designer/AGENT.md 2.3 KB
- agents/metrologist/AGENT.md 2.4 KB
- agents/process-modeler/AGENT.md 2.1 KB
- agents/quality-auditor/AGENT.md 2.5 KB
- agents/release-manager/AGENT.md 2.3 KB
- agents/verifier/AGENT.md 2.5 KB
- assessments/README.md 3.0 KB
- gates/G0-intake.md 1.5 KB
- gates/G1-requirements.md 2.3 KB
- gates/G2-design.md 2.1 KB
- gates/G3-interface.md 2.2 KB
- gates/G4-construction.md 2.6 KB
- gates/G5-verification.md 2.5 KB
- gates/G6-release.md 2.4 KB
- gates/G7-quality.md 2.9 KB
- gates/G8-closure.md 2.3 KB
- gates/GATES.md 4.0 KB
- instruments/cmmi-probe.md 3.6 KB
- instruments/INSTRUMENTS.md 2.0 KB
- instruments/metrics.md 4.4 KB
- instruments/quality-attributes.md 3.0 KB
- instruments/scripts/gate_check.py 11 KB runs code
- instruments/scripts/ledger.py 9.8 KB runs code
- instruments/scripts/metrics.py 11 KB runs code
- instruments/scripts/traceability.py 4.3 KB runs code
- interrogation/banks/01-requirements.md 3.2 KB
- interrogation/banks/02-design.md 2.8 KB
- interrogation/banks/03-interface.md 2.3 KB
- interrogation/banks/04-construction.md 2.8 KB
- interrogation/banks/05-verification.md 3.9 KB
- interrogation/banks/06-scm.md 2.9 KB
- interrogation/banks/07-quality.md 5.0 KB
- interrogation/banks/08-process.md 4.3 KB
- interrogation/banks/09-modeling.md 2.7 KB
- interrogation/banks/10-tooling.md 2.8 KB
- interrogation/banks/11-ethics-teams.md 6.2 KB
- interrogation/banks/scenarios.md 5.6 KB
- interrogation/intake.md 4.0 KB
- interrogation/QUESTIONS.md 2.2 KB
- kernel/constitution.md 5.5 KB
- kernel/glossary.md 8.3 KB
- kernel/KERNEL.md 6.5 KB
- kernel/ledger-spec.md 3.6 KB
- kernel/refusal-protocol.md 3.1 KB
- kernel/state-machine.md 2.7 KB
- manifest.json 2.5 KB
- modes/audit.md 4.1 KB
- modes/doctor.md 3.9 KB
- modes/drill.md 2.7 KB
- modes/fleet.md 5.1 KB
- modes/MODES.md 2.0 KB
- modes/task.md 3.8 KB
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 · 278 lines · 252 tokens per session scan A 889785f4844d
ilana is a skill published in the GitHub repository Ferousco-dev/ilana (8 stars, last pushed 7d ago), licensed MIT. It adds 252 tokens to every session and 3,324 once invoked, about $0.0013 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.
Other skills, from other repositories
biz-skill-creator
业务Skill的全生命周期管理：创建、修改、参数调整、测试验证。支持任意业务领域，首次接触新领域时自动从设计文档提取并积累领域知识。.
creating-skill
Builds and revises a complete skill DIRECTORY — SKILL.md, scripts, references, assets — and packages it. Use for "create a skill for X", "turn this into a skill", "update/improve this skill", "why doesn't my skill trigger", "review this SKILL.md", "package this skill", or when a repeated procedure should become a…
down-skilling
Distill Opus-level reasoning into optimized instructions for Haiku 4.5 (and Sonnet). Generates explicit, procedural prompts with n-shot examples that maximize smaller model performance on a given task. Use when user says "down-skill", "distill for Haiku", "optimize for Haiku", "make this work on Haiku", "generate…
agent-routing
Decide which model, effort level, and cascade shape each subagent gets, and how to keep improvement loops safe (evaluator-as-selector, stop on regression). Routes on measured cost-per-completed-task rather than per-token price, because a tier's token count varies more by task shape than price varies across tiers.…
challenging
Cross-context adversarial review for deliverables before shipping. Use when producing blog posts, technical recommendations, analysis briefs, code, or any artifact where accuracy matters more than speed. Triggers on "challenge this", "review before shipping", "adversarial pass", "stress test this".
crafting-instructions
Chooses the right FORMAT for instructions on Claude.ai — project instructions, a skill, or a standalone prompt — and gives the format-specific structure once chosen. Use when the question is which container an instruction belongs in ("should this be a skill or project instructions", "where do I put this", "how do I…