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/axiomantic/spellbookWrote 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/commands/axiomantic/spellbook/fractal-think-work)<a href="https://agentmods.dev/commands/axiomantic/spellbook/fractal-think-work"><img src="https://agentmods.dev/badge/commands/axiomantic/spellbook/fractal-think-work/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/commands/axiomantic/spellbook/fractal-think-work"><img src="https://agentmods.dev/badge/commands/axiomantic/spellbook/fractal-think-work.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 | $0.00022 | $0.04265 |
| Opus 5 | $0.00011 | $0.02132 |
| Sonnet 5 | $0.00004 | $0.00853 |
| Haiku 4.5 | $0.00002 | $0.00426 |
Grade A, and why
fractal-think-work 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 5d 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 — 464 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Phase 2: Fractal Think Work
Invariant Principles
- The graph IS the work queue - All state is in MCP tools. No separate task queue, no round counter, no cluster registry. Workers pull work dynamically via
fractal_claim_work. - All graph mutations happen inside workers - The dispatcher only reads state; it never writes nodes or synthesis.
- Return, don't re-dispatch - When workers exit with open work, return state to the orchestrator (SKILL.md). It decides whether to re-dispatch.
Before dispatching, assess: open question count, budget limits, graph active status. After workers finish, verify: no orphaned claimed nodes, synthesis cascade complete, all work accounted for.
Read skills/fractal-thinking/references/mcp-tools.md before Step 1, and pass that
path to every worker. It is the canonical definition of the fractal_* tool surface,
the valid intensity and checkpoint_mode values, the saturation reasons, and the node
state machine that governs every status transition workers make.
Parameters
| Parameter | Required | Description |
|---|---|---|
exploration_state |
Yes | JSON from Phase 1 (seed command) with graph_id, budget, intensity, checkpoint |
The exploration_state has this shape:
{
"graph_id": "<graph_id>",
"root_node_id": "<root_node_id>",
"intensity": "<intensity>",
"checkpoint": "<checkpoint>",
"budget": { "max_agents": 8, "max_depth": 4 },
"seed_count": 6
}
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.
- 5d ago First seen · 464 lines · 22 tokens per session scan A e654ef00e6ad
fractal-think-work is a command published in the GitHub repository axiomantic/spellbook (10 stars, last pushed yesterday), licensed MIT. It adds 22 tokens to every session and 4,265 once invoked, about $0.0001 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 2026-09-03.
Other commands, from other repositories
subagent-implementation
Orchestrate implement→review subagent loop until task complete. Reads the approved spec, writes a thin brief to .claude/.scratchpad/, dispatches fresh-context subagents, loops until reviewer signs off, commits per green iteration, then updates repo docs.
documentation
Bootstrap and maintain project documentation surfaces. Two modes: bootstrap (discover doc files, index them in CLAUDE.md) and authoring (scan for unindexed docs, match diff against indexed surfaces, walk stale/incomplete/missing items with Yes/Later/Remind/Skip).
watch-ci
Spawn a background Haiku-backed subagent to watch CI for the current branch (or specified target). Provider-agnostic — the subagent inspects project signals to identify the CI system (GitHub Actions, GitLab CI, CircleCI, etc.) and picks the right CLI. Returns immediately; reports back when CI reaches a terminal state.
session-report
Capture what changed this session and why, scoped to the current branch. Read by ship verbs when synthesizing the commit message; deleted after a successful commit.
integrate
Analyze and enhance AI artifacts to leverage Subcog memory effectively.
add-command
Add a new slash command to the current plugin.