construct-counterexamples

A reasoning aid for testing a proposed claim by looking for examples where its assumptions hold but its conclusion fails. Such an example is called a counterexample.

In plain words
What is it for?
Use it to challenge conjectures, lemmas, or intermediate reasoning claims and record any examples that refute them.
Why use it?
It helps reveal false claims and weak proof steps before they are treated as correct. If no counterexample is found, that only means the claim has not yet been disproved.

Skill for Claude CodeCodex

Install

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.

agentmods
npx agentmods add skills/frenzymath/danus/construct-counterexamples
Any agent
npx skills add frenzymath/Danus --skill construct-counterexamples
Clone the repo
git clone --depth 1 https://github.com/frenzymath/Danus

Made for: Claude Code, Codex.

Per session 66 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 711 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00066 $0.00711
Opus 5 $0.00033 $0.00356
Sonnet 5 $0.00013 $0.00142
Haiku 4.5 $0.00007 $0.00071

Measured 3d ago against content hash dac7c1207a04, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

construct-counterexamples 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 3d 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.

agents/skills/worker/construct-counterexamples/SKILL.md · 82 lines

How it starts

The opening of the file, as written. The whole thing — 82 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Construct Counterexamples

Actively falsify proposed conjectures or intermediate claims by finding examples that satisfy the assumptions but violate the claimed conclusion.

Input Contract

Read:

  • the specific conjecture/claim to test
  • active branch assumptions
  • candidate lemmas/proof steps
  • current immediate_conclusions and toy_examples
  • previously found counterexamples that can be reused against new claims

Procedure

  1. Identify the assumptions that must hold and the conclusion to fail.
  2. Use reasoning, decomposition, and retrieval to search for standard obstructions, pathological constructions, or previously known counterexamples.
  3. Decide status:
    • refuted: assumptions hold and the claim fails
    • not_refuted: no counterexample found yet
    • inconclusive: search space unclear or partially explored
  4. If the search produces a concrete example that is informative but is not actually a counterexample, save that example as well in toy_examples.
  5. If refuted, store the counterexample for reuse against future claims and mark impacted branches/lemmas as invalid.
  6. If no counterexample is found, treat that only as evidence that the claim may be correct, not as a proof.

Output Contract

Publish to global memory with gm_add (kind counterexample): claim = what is refuted/tested, evidence = the candidate construction, plus these fields:

{
  "target_claim": "...",
  "candidate_counterexample": "...",
  "status": "refuted|not_refuted|inconclusive",
  "assumptions_satisfied": ["..."],
  "failed_conclusion": "...",
  "impact": "...",
  "branch_id": "optional",
  "subgoal_id": "optional"
}

If status="refuted" and it kills a branch, also publish a dead_end finding (gm_add, kind dead_end) so siblings skip that branch.

If the search produced a concrete non-refuting example, also publish an example finding (gm_add, kind example):

{
  "example": "...",
  "why_relevant": "constructed while testing the claim ...",
  "assumptions_satisfied": ["..."],
  "conclusion_verified": true,
  "where_assumptions_take_effect": "...",
  "observed_pattern": "...",
  "supports_branch_ids": ["optional"],
  "subgoal_id": "optional"
}

Read the full file on GitHub · 82 lines

Files

What ships with it

1 file 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.

Changes

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.

  1. 3d ago First seen · 82 lines · 66 tokens per session scan A dac7c1207a04

Subscribe to this mod's changes

construct-counterexamples is a skill published in the GitHub repository frenzymath/Danus (387 stars, last pushed 7d ago), licensed Apache-2.0. It adds 66 tokens to every session and 711 once invoked, about $0.0003 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.

Related

Other skills, from other repositories

agent-matrix-optimizer

Agent skill for matrix-optimizer - invoke with $agent-matrix-optimizer.

ruvnet/ruflo · 21 tokens

hugging-face-trackio

Track and visualize ML training experiments with Trackio. Use when logging metrics during training (Python API) or retrieving/analyzing logged metrics (CLI). Supports real-time dashboard visualization, HF Space syncing, and JSON output for automation.

patchy631/ai-engineering-hub · 52 tokens

papyrus-writing

LaTeX paper writing and editing. Load when working on a .tex document — writing or revising sections, fixing compilation errors, adding figures/tables/equations, or managing a bibliography. Also the co-author skill for the Papyrus app.

kirodotdev/KiroCrew · 55 tokens

arithmetic-evaluator

Evaluate arithmetic expressions and return numeric results. Use when a user asks an arithmetic question, requests calculation, or provides a math expression involving numbers, parentheses, and arithmetic operators.

JetBrains/koog · 39 tokens

tooluniverse-cell-line-profiling

Cancer cell-line selection and profiling for experimental model choice. Cross-references DepMap, Cellosaurus, COSMIC, PharmacoDB to deliver identity verification, mutation/CNV profile, gene dependencies, drug sensitivities, and druggable targets. Use to answer 'which cell line should I use for studying gene X?' or 'is…

mims-harvard/ToolUniverse · 100 tokens

tooluniverse-computational-biophysics

Solve quantitative problems in biophysics — pharmacokinetics (PK volume of distribution, clearance, half-life), epidemiology (R0, attack rate), toxicology (LD50, NOAEL), population genetics (Hardy-Weinberg, Fst), enzyme kinetics (Michaelis-Menten), thermodynamics. Use for first-principles quantitative biology…

mims-harvard/ToolUniverse · 97 tokens