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 skills add WdBlink/autoresearch-paper --skill autoresearch-workflowgit clone --depth 1 https://github.com/WdBlink/autoresearch-paperWrote 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/skills/wdblink/autoresearch-paper/autoresearch-workflow)<a href="https://agentmods.dev/skills/wdblink/autoresearch-paper/autoresearch-workflow"><img src="https://agentmods.dev/badge/skills/wdblink/autoresearch-paper/autoresearch-workflow/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/skills/wdblink/autoresearch-paper/autoresearch-workflow"><img src="https://agentmods.dev/badge/skills/wdblink/autoresearch-paper/autoresearch-workflow.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 | $0.00044 | $0.00724 |
| Opus 5.5 | $0.00018 | $0.00290 |
| Sonnet 5.5 | $0.00009 | $0.00145 |
| Haiku 4.5 | $0.00004 | $0.00072 |
Grade A, and why
autoresearch-workflow 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 4d 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 — 77 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Core contract
Act only as a thin router. Accept either an explicit entry request or one four-field compact handoff, select the first matching state in the canonical route matrix, emit exactly one four-field handoff, and stop. Never inspect repository internals, load another skill, create domain artifacts, or perform research work.
Input
The input handoff has next_skill, reason, input_artifact, and
resume_artifact. Its reason begins status=<canonical-token>;. For direct
entry, use only the artifacts, status, and human confirmation explicitly stated;
treat a referenced-but-missing required input as missing rather than inferring a
different lifecycle state.
Read artifact handoffs to obtain the ordered route matrix. Negative, refusal, confirmation, evaluator-return, and terminal rows precede artifact-presence fallthrough. Select the first matching row.
Routing
The only domain destinations are autoresearch-discovery,
karpathy-autoresearch-adapter, autoresearch-evaluator-engineering,
autoresearch-radar, autoresearch-experiment, autoresearch-evidence, and
autoresearch-paper.
Use literal none for a terminal outcome, confirmation-pending state, refused
package, or any other canonical no-route row. Do not route back to Discovery
after a Research Brief exists.
Adapter's repository-not-runnable and baseline-failed outcomes are terminal,
use next_skill: none, and precede the Research-Brief fallback. Experiment
evaluator failure routes the bound autoresearch/evaluator-invalid-return.md
to Adapter, never the stale contract.
Evaluator construction is a conditional operational detour: Adapter sends a
partial or missing evaluator to Evaluator Engineering; an Evaluator Package
returns to Adapter. An evaluator integrity/readiness problem found by Experiment
also returns to Adapter. Neither path is a scientific return loop.
Research Radar is a conditional operational sidecar, never a lifecycle stage.
Route an explicit research-radar-refresh-requested to it using only the
Adapter-issued autoresearch/research-watch-profile.yaml. Its complete,
sync-failed, cycle-failed, and commit-indeterminate outcomes all route to
none; the external scheduler or a later explicit request owns any future
scan. An indeterminate commit must be inspected before another refresh is
requested.
What ships with it
2 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.
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.
- 4d ago First seen · 77 lines · 44 tokens per session scan A 03b08aadcd71
autoresearch-workflow is a skill published in the GitHub repository WdBlink/autoresearch-paper (2 stars, last pushed 2d ago), licensed MIT. It adds 44 tokens to every session and 724 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-11-02.
autoresearch-workflow compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| autoresearch-workflowWdBlink/autoresearch-paper | Skill | 2 | 2d ago | MIT | A | 44 | Codex |
| data-management-planpedrohcgs/claude-code-my-workflow | Skill | 1.7k | 2d ago | MIT | A | 143 | Claude Code |
| scaffold-exercisespedrohcgs/claude-code-my-workflow | Skill | 1.7k | 2d ago | MIT | A | 101 | Claude Code |
| validate-bibpedrohcgs/claude-code-my-workflow | Skill | 1.7k | 2d ago | MIT | A | 47 | Claude Code |
| anti-defensive-writing-enAdkid-Zephyr/anti-defensive-writing-Skill | Skill | 5.3k | 10d ago | MIT | A | 138 | Claude Code, Codex |
| flux-analyzeraiming-lab/AutoResearchClaw | Skill | 15k | 2mo ago | MIT | A | 44 | Claude Code, Codex |
| gsmm-builderaiming-lab/AutoResearchClaw | Skill | 15k | 2mo ago | MIT | A | 45 | Claude Code, Codex |