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 orneryd/NornicDB --skill decay-tuninggit clone --depth 1 https://github.com/orneryd/NornicDBWrote 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/orneryd/nornicdb/decay-tuning)<a href="https://agentmods.dev/skills/orneryd/nornicdb/decay-tuning"><img src="https://agentmods.dev/badge/skills/orneryd/nornicdb/decay-tuning/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/orneryd/nornicdb/decay-tuning"><img src="https://agentmods.dev/badge/skills/orneryd/nornicdb/decay-tuning.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 | $0.00082 | $0.02358 |
| Opus 5 | $0.00041 | $0.01179 |
| Sonnet 5 | $0.00016 | $0.00472 |
| Haiku 4.5 | $0.00008 | $0.00236 |
Grade A, and why
nornicdb-decay-tuning 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 — 175 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Tuning NornicDB Decay Profiles
This skill is the playbook for picking and adjusting decay parameters. It assumes the vocabulary from the nornicdb-knowledge-policies skill: a bundle is a parameter set (CREATE DECAY PROFILE <name> OPTIONS { ... }, no target, inert), a binding attaches decay to entities (CREATE DECAY PROFILE <name> FOR (...) APPLY { ... }).
All tuning is done by either editing a bundle (ALTER DECAY PROFILE <bundle> SET OPTIONS { ... }) or by setting overrides inside a binding's APPLY block. Decay runs on every read regardless of whether any promotion policy exists — ON ACCESS is a promotion concern, not a decay one.
The four levers
| Lever | What it controls | Independent of |
|---|---|---|
halfLifeSeconds |
Time to fall to 0.5 (negative inverts the curve) | All others |
function |
Curve shape: exponential / linear / step / none |
Half-life value |
visibilityThreshold |
Boolean cutoff. finalScore < threshold ⇒ entity hidden |
Score itself |
scoreFloor |
max() clamp on the reported score |
Visibility |
scoreFloor and visibilityThreshold are the two parameters most people misuse. A floor only makes something visible if scoreFloor >= visibilityThreshold. Otherwise the floor just keeps the score off zero while it stays suppressed.
Picking halfLifeSeconds
Start from "how long should this still be visible by default?" and divide by ~3.32 to get the threshold-crossing time at the default 0.10 threshold:
| Half-life | Crosses 0.10 (exp) at | Use for |
|---|---|---|
| 3600 s (1h) | ~3.3h | ephemeral working state, scratch |
| 86400 s (1d) | ~3.3d | sessions, short-lived signals |
| 604800 s (1w) | ~23d | typical document/episode memory |
| 2592000 s (30d) | ~3.3 months | reference material that ages slowly |
none / NO DECAY |
never | identifiers, canonical links |
For linear the same threshold is reached at (1 - threshold) * 2 * halfLife. For step everything is full-score until the half-life, then zero — useful only when "expires sharply" is the model.
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 · 175 lines · 82 tokens per session scan A 0490cfcd65a7
nornicdb-decay-tuning is a skill published in the GitHub repository orneryd/NornicDB (864 stars, last pushed yesterday), licensed MIT. It adds 82 tokens to every session and 2,358 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 2026-09-05.
Other skills, from other repositories
payload
Use when working with Payload projects (payload.config.ts, collections, fields, hooks, access control, Payload API). Use when debugging validation errors, security issues, relationship queries, transactions, or hook behavior.
graphjin-eval
Create, extend, run, baseline, and diagnose GraphJin agent evaluations through the graphjin eval CLI.
graphjin-env
Use when setting up a training or evaluation loop against a GraphJin agent environment — running the container, reading /health, driving episodes hosted or step-by-step or with your own agent over MCP, splitting train from eval, exporting trajectories, and deciding whether two rewards can be compared.
infrahub-sot
OpsMill Infrahub — infrastructure source of truth with schema-driven nodes, GraphQL queries, and branch-isolated changes. Use when querying Infrahub for device/IPAM inventory, browsing infrastructure schemas, running GraphQL queries, or making infrastructure changes safely via an auto-created session branch and a…
dkg-importer
Bulk-import a large RDF graph (code graph, corpus, GitHub history, etc.) into a DKG node's working memory. Use this skill when you need to push more than a few thousand triples in a single import — it codifies the chunking budgets, the assertion-loop shape, the resumability manifest, and the canonical URI rules so…
pydantic-resolve-3step
A three-stage development method for Python projects built with pydantic-resolve, starting with agreed data models and ending with application interfaces. It uses an ER diagram, which shows database entities and their relationships, plus ORM models that represent database records in code.