Borrowing it
Nothing to install: this file belongs to TheSmokeDev/taskchad-os. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/TheSmokeDev/taskchad-os/master/.claude/skills/ai-citation-authority-wave/assets/archon/commands/citation-authority-research.mdgit clone --depth 1 https://github.com/TheSmokeDev/taskchad-osWrote 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/thesmokedev/taskchad-os/citation-authority-research)<a href="https://agentmods.dev/commands/thesmokedev/taskchad-os/citation-authority-research"><img src="https://agentmods.dev/badge/commands/thesmokedev/taskchad-os/citation-authority-research/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/thesmokedev/taskchad-os/citation-authority-research"><img src="https://agentmods.dev/badge/commands/thesmokedev/taskchad-os/citation-authority-research.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 | $0.00023 | $0.00689 |
| Opus 5 | $0.00012 | $0.00345 |
| Sonnet 5 | $0.00005 | $0.00138 |
| Haiku 4.5 | $0.00002 | $0.00069 |
Grade A, and why
citation-authority-research 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 11d 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 — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Research one bounded authority wave
Work only in the current website repository. This node may research and write
.citation-authority/evidence-packet.json. It must not edit site content, use a
Reddit account, publish, deploy, submit indexing requests, or manufacture a
metric.
Inputs
Read these files in order:
.citation-authority/run-config.json.citation-authority/site-profile.json.citation-authority/evidence-request.json, when present.citation-authority/evidence-packet.json, when already present- the optional fleet intent map named by
run-config.json - existing content near the configured content sink
If evidence-packet.json already has status: ready or status: no_evidence,
preserve it byte for byte and report EVIDENCE_PRESERVED.
Receipt hierarchy
Prefer existing first-party receipts in this order:
- GSC queries with impressions or clicks and a stated date range.
- OpenSEO measurements with a run ID or source URL.
- A live SERP autopsy performed in this node.
For a SERP autopsy, search the exact candidate query in the site's configured language and record at least three URLs that were actually returned. Capture the engine, observation timestamp, titles/positions when visible, the buyer intent, and the specific gap. A Reddit-modifier receipt should include a real Reddit discussion URL when one ranks. Merely finding a Reddit thread is not a claim that Reddit recommends this brand.
Do not convert keyword-tool silence into a made-up zero. No data is a valid
note, not a numeric metric. Do not infer GSC data from a public SERP. Do not use
stale receipts older than 90 days.
Language and fleet boundaries
entargets are researched and written as English-native intents.estargets are researched as Spanish-native intents, not translations of an English list. Use natural terms such ascotizacion, not Englishquote.- Do not put Spanish pages under
/eson a Spanish-only root-domain profile. - Respect the fleet intent map. A query already owned by another domain is not eligible here.
- Keep the research broad enough to compare Reddit modifiers, direct
who helps...answers, and comparison intents, but do not select targets in this node.
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.
- 11d ago First seen · 80 lines · 23 tokens per session scan A 4ae0d7e3abda
citation-authority-research is a command published in the GitHub repository TheSmokeDev/taskchad-os (23 stars, last pushed today), licensed MIT. It adds 23 tokens to every session and 689 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 2026-08-30.
Other commands, from other repositories
second-brain-mapping
Map your vault: extract structured metadata from every typed file, surface cross-doc insights, optionally build a knowledge graph.
setup-vault-types
Configure which document types your vault uses (journals, books, meetings, clients, etc.) and scaffold extractors.
cierre
A sales call just ended: turn its transcript into the full follow-up (CRM, tasks, email draft, reminder, coaching).
daily-journal
Daily journal interview and entry creator with emotional floor tagging.
deconstruct
First-principles analyst: surface hidden assumptions, find foundational truths, rebuild from scratch.
diagnose
Run a self-check on your AI Brain Starter install (CLAUDE.md, Meta folder, skills, hooks, MCPs).