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/fact-check-verify)<a href="https://agentmods.dev/commands/axiomantic/spellbook/fact-check-verify"><img src="https://agentmods.dev/badge/commands/axiomantic/spellbook/fact-check-verify/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/fact-check-verify"><img src="https://agentmods.dev/badge/commands/axiomantic/spellbook/fact-check-verify.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 | $0.00016 | $0.01564 |
| Opus 5 | $0.00008 | $0.00782 |
| Sonnet 5 | $0.00003 | $0.00313 |
| Haiku 4.5 | $0.00002 | $0.00156 |
Grade A, and why
fact-check-verify 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 — 131 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Shared Reference: This command uses the evidence hierarchy and depth escalation protocol defined in
skills/shared-references/evidence-hierarchy.md.
Fact-Check: Parallel Verification and Verdicts (Phases 4-5)
Invariant Principles
- Check the run checkpoint before re-verifying - Read
.fact-checking/state.json(the current run's resume state) before re-verifying a claim; skip already-completed claims on resume and use the recorded verdict when nothing has changed in scope. - Evidence requires source - Every verdict must cite code traces, test results, docs, or benchmarks.
- Cross-agent confirmation - A claim verified by one agent is not confirmed until a second independent agent corroborates or no contradicting evidence emerges.
Phase 4: Parallel Verification
Read the run checkpoint BEFORE verifying. Write the run checkpoint AFTER each completed claim.
The checkpoint lives at .fact-checking/state.json (project-local, gitignored).
It is the run's resume artifact, not a cross-run knowledge base. The read/write
contract:
- Before verifying a claim: read the checkpoint. If the claim is in the
completedlist and the file referenced has not changed since the recorded timestamp (compare againstgit statusor file mtime), reuse the recorded verdict and bibliography. If the file has changed, drop the cached entry and re-verify. Cross-run knowledge is intentionally out of scope. - After verifying a claim: write the verdict, bibliography entry, and
timestamp into the checkpoint's
findingsmap and append the claim index tocompleted. The checkpoint is the source of truth for resume; without it, a resumed run will redo work that has already been verified.
Error paths:
- Checkpoint file missing or corrupt: treat as fresh run; do not block verification.
- Checkpoint write fails (disk full, permissions): log and continue; verification still completes for this claim, but the run is no longer resumable from this point.
- Swarm spawn failure: escalate to orchestrator; do not issue partial verdicts.
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 · 131 lines · 16 tokens per session scan A fc13ae31a7a1
fact-check-verify is a command published in the GitHub repository axiomantic/spellbook (10 stars, last pushed yesterday), licensed MIT. It adds 16 tokens to every session and 1,564 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
review-branch
Review the current branch's diff against base by dispatching atomic-reviewer. No orchestration loop, no spec required — pre-flight before /commit pr or /commit merge.
dream
Overnight autoresearch + maintainer sweep that surfaces a MORNING REPORT — suggested changes, issues raised with fixes, and kept improvements. Shadow-first; never auto-pushes. Runs on-demand or scheduled for off-hours.
code-quality-plan-creator
LSP-powered architectural code quality analysis - works with any executor (loop or swarm).
mr-description-creator
Generate and apply MR/PR description directly via gh or glab CLI (project).
follow-up
Review and act on pending reminders. Bare invocation shows all reminders as an indexed list; cron-fired invocation (/follow-up due ) surfaces the specific reminder and waits for a response; /follow-up review lists stale project follow-up entries for per-item disposition (extend/close/promote/skip). Transport-aware…
pr-review
Generate a PR review report aggregating quality scan, coverage, complexity, and breaking changes.