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 benjaminard/fable-skills --skill evidence-audited-analysisgit clone --depth 1 https://github.com/benjaminard/fable-skillsWrote 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/benjaminard/fable-skills/evidence-audited-analysis)<a href="https://agentmods.dev/skills/benjaminard/fable-skills/evidence-audited-analysis"><img src="https://agentmods.dev/badge/skills/benjaminard/fable-skills/evidence-audited-analysis/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/benjaminard/fable-skills/evidence-audited-analysis"><img src="https://agentmods.dev/badge/skills/benjaminard/fable-skills/evidence-audited-analysis.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 | $0.00068 | $0.00702 |
| Opus 5 | $0.00034 | $0.00351 |
| Sonnet 5 | $0.00014 | $0.00140 |
| Haiku 4.5 | $0.00007 | $0.00070 |
Grade A, and why
evidence-audited-analysis 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 10d 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 — 35 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Evidence-Audited Analysis
The most common analytical failure is not a wrong model. It is a confident, well-written narrative built on numbers nobody interrogated. Audit the evidence before you analyze it, and state what it cannot show alongside what it can.
Interrogate the data first
Before any analysis, profile what you were given and report anomalies:
- Shape: row counts, date coverage, obvious gaps. An analysis of "last quarter" on a table that only starts mid-quarter is wrong before it begins.
- Quality: nulls, duplicates, impossible values (negative durations, future timestamps), unit surprises (cents vs dollars, UTC vs local).
- Meaning: for every field you lean on, answer in writing: what does this number actually measure, and who put it there? A "deal created" count measures logging behavior, not demand. A "page view" count includes bots unless someone excluded them. If a human process feeds the field, the field inherits that process's habits, and you must ask about them before treating the field as ground truth.
If a profiling step surfaces something odd, resolve it or disclose it. Never silently analyze around it.
Audit your own numbers
- Reproduce every headline number a second, independent way before reporting it: a different query, a different aggregation path, or a manual spot-check of raw rows. If two routes disagree, that is a finding about the data, and it comes before any other finding.
- Baseline before model. Report the naive answer (last period's value, the overall average, the simplest split) before anything sophisticated. If the sophisticated answer does not beat the baseline meaningfully, say so; the baseline is the finding.
- Small samples get small claims. State the n behind every rate or comparison. "Conversion doubled" on 4 versus 2 events is noise wearing a trend's clothes. When the sample cannot support the conclusion, the honest deliverable is "this data cannot answer that yet," with what would be needed.
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.
- 10d ago First seen · 35 lines · 0 tokens per session scan A 6b537dc9ef69
evidence-audited-analysis is a skill published in the GitHub repository benjaminard/fable-skills (30 stars, last pushed 2mo ago), licensed MIT. It adds 68 tokens to every session and 702 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.
Other skills, from other repositories
guidance
Control LLM output with regex and grammars, guarantee valid JSON/XML/code generation, enforce structured formats, and build multi-step workflows with Guidance - Microsoft Research's constrained generation framework.
alterlab-paper-writer
Drafts and revises academic papers through a 12-agent pipeline with hardened LaTeX output (apa7 document class, justified text, table column-width formula, centered bilingual abstracts, standardized font stack, PDF compiled from LaTeX), supporting IMRaD, literature review, theoretical, case study, policy brief, and…
alterlab-pyhealth
Develops, tests, and deploys clinical machine learning models with the PyHealth healthcare AI toolkit. Use when working with electronic health records (EHR), clinical prediction tasks (mortality, readmission, drug recommendation), medical coding systems (ICD, NDC, ATC), physiological signals (EEG, ECG), healthcare…
alterlab-imaging-data-commons
Query and download public cancer imaging data from the NCI Imaging Data Commons (IDC) using the idc-index Python package, filtering by metadata, visualizing in-browser, and checking licenses, with no authentication required. Use when obtaining large-scale radiology (CT, MR, PET) or digital pathology DICOM datasets for…
alterlab-cobrapy
Build and analyze genome-scale constraint-based metabolic models with COBRApy — flux balance analysis (FBA), flux variability analysis (FVA), gene and reaction knockouts, flux sampling, and SBML model I/O. Use when simulating metabolic networks, predicting growth or knockout phenotypes, or running systems-biology and…
alterlab-phylogenetics
Build phylogenetic trees end-to-end from raw sequences — MAFFT multiple sequence alignment, optional TrimAl trimming, IQ-TREE 2 maximum-likelihood inference with model selection and bootstraps, FastTree for large datasets, then visualize with ETE3 or FigTree. Use when reconstructing trees from sequences (FASTA) for…