Borrowing it
Nothing to install: this file belongs to cpuguy96/StepCOVNet. 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/cpuguy96/StepCOVNet/master/.cursor/skills/tide-ablations/SKILL.mdgit clone --depth 1 https://github.com/cpuguy96/StepCOVNetWrote 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/cpuguy96/stepcovnet/tide-ablations)<a href="https://agentmods.dev/skills/cpuguy96/stepcovnet/tide-ablations"><img src="https://agentmods.dev/badge/skills/cpuguy96/stepcovnet/tide-ablations/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/cpuguy96/stepcovnet/tide-ablations"><img src="https://agentmods.dev/badge/skills/cpuguy96/stepcovnet/tide-ablations.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 | $0.00053 | $0.00585 |
| Opus 5 | $0.00026 | $0.00293 |
| Sonnet 5 | $0.00011 | $0.00117 |
| Haiku 4.5 | $0.00005 | $0.00059 |
Grade A, and why
tide-ablations 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 — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Tide ablations
Script
scripts/run_overfit_tide_ablations.py
Use the WSL command template with:
scripts/run_overfit_tide_ablations.py --epochs=50 --phase=all
Phases:
| Phase | What it does |
|---|---|
threshold |
Sweep confidence on existing MERT checkpoint (no retrain) |
train |
Retrain MERT with loss-weight and architecture variants |
all |
Threshold sweep then training ablations (default) |
Output
| Artifact | Path |
|---|---|
| Summary | models_wsl/overfit_tide_ablations/ablation_summary.json |
| Per-run models | models_wsl/overfit_tide_ablations/<variant>/ |
| Callbacks | callbacks/overfit_tide_ablations/<variant>/ |
EXP-10 conclusions (do not re-run blindly)
- Threshold sweep: no meaningful F1 lift — recall capped before post-hoc threshold helps
- Loss weights (cls-heavy, time-heavy, both-high): ~same ~28–30% F1
- arch_medium (embed 128): similar to baseline
- arch_large (embed 256): OOM on 8 GB GPU — skip or reduce batch
Plateau is not fixed by epochs, loss weights, threshold, or medium arch bump. Likely query-slot formulation (high FP, ~233/634 recall).
After the run
- Read
ablation_summary.json(if script crashed, check terminal + partial models) - Log
EXP-…in EXPERIMENT_LOG.md; tag pipeline stages:trainfor loss,modelfor arch,postfor threshold - Next: multi-song val or formulation alternatives per EXPERIMENT_LOG.md § Recommended next step
Known pitfall
sweep_confidence_thresholds requires numeric min_onset_distance_ms (use 0.0, not None). See JRN-20260606-05 in self-journal.md.
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 · 51 lines · 53 tokens per session scan A 42d133463ffc
tide-ablations is a skill published in the GitHub repository cpuguy96/StepCOVNet (22 stars, last pushed 15d ago), licensed Apache-2.0. It adds 53 tokens to every session and 585 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
physicsnemo-discover
Official NVIDIA-authored guidance for navigating PhysicsNeMo — pick the model, datapipe, or example for a SciML/AI4Science task (surrogates, forecasting, downscaling, physics-informed, inverse, generative). Points at existing files via live repo search; never writes code. Do NOT use for installation or environment…
ml-for-aec
Computer vision for buildings, image-to-floorplan, generative ML models, performance prediction, structural analysis ML, energy prediction, natural language to design, and point cloud ML for AEC computational design.
ml-architecture-diagram
Create accurate, publication-ready, editable machine-learning and deep-learning architecture diagrams from model code, configuration files, model summaries, graph exports, or written specifications. Use for neural-network figures, architecture schematics, model block diagrams, training/inference diagrams, paper…
science-vibecoding
Structured AI-assisted scientific code generation. 6 safety guards, 8 principles, 11 prompt templates. Grounded in Nature (2026).
ml-for-aec
Computer vision for buildings, image-to-floorplan, generative ML models, performance prediction, structural analysis ML, energy prediction, natural language to design, and point cloud ML for AEC computational design.
flow-nexus-neural
Train and deploy neural networks in distributed E2B sandboxes with Flow Nexus.