Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/zjunlp/Mechanistnpx agentmods add skills/zjunlp/mechanist/pick-alternativesWrote 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/zjunlp/mechanist/pick-alternatives)<a href="https://agentmods.dev/skills/zjunlp/mechanist/pick-alternatives"><img src="https://agentmods.dev/badge/skills/zjunlp/mechanist/pick-alternatives/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/zjunlp/mechanist/pick-alternatives"><img src="https://agentmods.dev/badge/skills/zjunlp/mechanist/pick-alternatives.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Agent Snooping · line 52 Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.Fix: Remove all code or instructions that list or read other skills' files or directories. Skills should operate independently; cross-skill access is a privilege escalation.
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 | $0.00081 | $0.03983 |
| Opus 5 | $0.00041 | $0.01992 |
| Sonnet 5 | $0.00016 | $0.00797 |
| Haiku 4.5 | $0.00008 | $0.00398 |
Grade A, and why
verify-pick-alternatives 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 9d 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 — 232 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Verify: Pick Alternatives
Sub-skill of /auto-verify. Chooses the variants that will test claim robustness along three independent dimensions: method, dataset, model.
For: $ARGUMENTS
Purpose
/auto-verify needs swap candidates that genuinely stress the claim — not cosmetic re-runs. This skill:
- Harvests candidates from existing project research (idea-stage, refine-logs, findings.md, research-wiki if present)
- Checks each dimension (method / dataset / model) for a minimum coverage threshold
- Calls
/research-litonly for dimensions below the threshold — a focused top-up, not a full survey - Hands the candidate pool to the external LLM reviewer, who picks and ranks the variants by "strongest independent test of the claim"
- Emits a structured variant list that
/auto-verifydrops intoverify/<claim_dir>/PLAN.md
Constants
-
DIMENSIONS =
method,dataset,model— Comma-separated subset of{method, dataset, model}, passed through from/auto-verify. Dimensions not in the list get zero variants — no candidate harvesting, no reviewer prompts for them. Parse from$ARGUMENTStail: the invoker passes— dimensions: method,dataset. If the parameter is absent, default to all three axes.Variant count =
len(DIMENSIONS), exactly. One swap per listed axis, no scaling. There is no separate "effort" knob; deeper analyses (multi-seed within an axis, 2-factor cross-axis interactions) belong in/ablation-planner, not here. -
MIN_CANDIDATES_PER_DIMENSION = 2 — if fewer than 2 credible candidates exist in any ACTIVE dimension, invoke
/research-litfor that dimension. Inactive dimensions are skipped entirely. -
RESEARCH_LIT_SCOPE =
focused— when invoked,/research-litruns in focused mode (≤ 6 papers, single dimension) to avoid a full-depth survey. -
REVIEWER_BACKEND =
llm-chat— External LLM reviewer via llm-chat MCP. Always ask for strict, high-rigor feedback.
Inputs
Parse $ARGUMENTS as [claim-id] — [claim-statement] [— dimensions: <subset>]. The trailing — dimensions: clause is optional; if present, split on ,, lowercase each entry, validate against {method, dataset, model}, reject unknown axes, and store as the active dimension list. If absent, the active list defaults to all three axes.
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.
- 9d ago First seen · 232 lines · 81 tokens per session scan A 14d4f8e2c618
verify-pick-alternatives is a skill published in the GitHub repository zjunlp/Mechanist (74 stars, last pushed 13d ago), licensed MIT. It adds 81 tokens to every session and 3,983 once invoked, about $0.0004 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 skills, from other repositories
nanoresearch-writing
Draft a LaTeX research paper from all previous stage outputs.
nanoresearch-experiment
Generate a Python code skeleton from an experiment blueprint.
nanoresearch-planning
Produce an experiment blueprint from a research hypothesis.
nanoresearch-ideation
Search academic literature and generate research hypotheses.
autoresearch
Orchestrates end-to-end autonomous AI research projects using a two-loop architecture. The inner loop runs rapid experiment iterations with clear optimization targets. The outer loop synthesizes results, identifies patterns, and steers research direction. Routes to domain-specific skills for execution, supports…
skypilot-multi-cloud-orchestration
Multi-cloud orchestration for ML workloads with automatic cost optimization. Use when you need to run training or batch jobs across multiple clouds, leverage spot instances with auto-recovery, or optimize GPU costs across providers.