Borrowing it
Nothing to install: this file belongs to FilipNowakowicz/nixos-config. 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/FilipNowakowicz/nixos-config/main/.agents/skills/capture-learning-candidate/SKILL.mdgit clone --depth 1 https://github.com/FilipNowakowicz/nixos-configWrote 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/filipnowakowicz/nixos-config/capture-learning-candidate)<a href="https://agentmods.dev/skills/filipnowakowicz/nixos-config/capture-learning-candidate"><img src="https://agentmods.dev/badge/skills/filipnowakowicz/nixos-config/capture-learning-candidate/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/filipnowakowicz/nixos-config/capture-learning-candidate"><img src="https://agentmods.dev/badge/skills/filipnowakowicz/nixos-config/capture-learning-candidate.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 | $0.00042 | $0.00763 |
| Opus 5.5 | $0.00017 | $0.00305 |
| Sonnet 5.5 | $0.00008 | $0.00153 |
| Haiku 4.5 | $0.00004 | $0.00076 |
Grade A, and why
capture-learning-candidate 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 1mo 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 — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Capture Learning Candidate
When work surfaces something a future agent should know, record one compact
candidate under .agents/learning/candidates/. Candidates are machine-routed
proposals reviewed later by a human-gated reviewer — see
.agents/learning/README.md. Capturing a candidate must never change agent
behavior on its own.
When to capture (high bar)
File a candidate only when all hold:
- The lesson is reusable across future sessions, not a one-off fact.
- You have verifiable evidence (commit,
file:line, failing check output, or a direct user correction). - It is not already encoded in
CLAUDE.md,lib/invariants.nix, a test, a hook, or an existing skill.
Strong triggers: a user correction of how you worked; a validation gate
(scripts/validate.sh, merge-gate) catching something a careful agent would
have avoided; a non-obvious repo gotcha you only learned by hitting it; or a
repo/CI fix that should become a durable check, hook, skill, or doc update.
When NOT to capture
- One-off or session-only context (belongs in your normal summary, nowhere else).
- Anything already covered by docs, invariants, tests, hooks, or skills.
- Ephemeral user/preference facts better suited to auto-memory.
- A lesson you cannot back with evidence.
Prefer fewer, higher-signal candidates. Skipping is the default.
Workflow
- Do not scan candidate bodies. Build 3-6 query terms from the issue shape, file paths, commands, CI job, host, or hook involved.
- Run
bash .agents/learning/scripts/query-candidates.sh <terms>.- If it returns a likely duplicate, open only that candidate file and update it in place.
- If it returns nothing relevant, create a new candidate.
- Copy
.agents/learning/TEMPLATE.ymlto.agents/learning/candidates/<date>-<kebab-slug>.yml. - Fill the required fields with terse, grep-friendly values:
route:implement-fix,promote-memory,promote-skill,promote-hook,promote-doc, orreject.best_form: strongest viable artifact, preferring executable checks over hooks, hooks over skills, and skills over prose docs.evidence,observation,proposed_upgrade, plusdate/expires/status. The optional fields (triggers,targets,dedupe_key,type,risk,agent) are commented out in the template — add one only when it adds real routing signal. Do not pad fields to look thorough.
- Mention the captured candidate's path in your work summary.
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.
- 1mo ago First seen · 68 lines · 42 tokens per session scan A 22b12e8bff68
capture-learning-candidate is a skill published in the GitHub repository FilipNowakowicz/nixos-config (5 stars, last pushed 6d ago), licensed MIT. It adds 42 tokens to every session and 763 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-08-31.
Other skills, from other repositories
agent-self-maintenance
A set of checks for keeping an AI coding assistant's journal, facts, observations, and self-check results current and clearly supported by evidence.
knowledge-ingest
Ingest documents, articles, conversations, or supplied text into Knowledge with retained originals, sourced claims, entity updates, timelines, and cross-links; also learn durable lessons from a set of conversations, transcripts, or past sessions (/learn), including the Learn from conversations and Learn from Memory…
mem0-oss-to-platform
Plan, then execute, a migration of a project from the mem0 open-source / self-hosted SDK (local Memory class) to the mem0 Platform / hosted SDK (MemoryClient). TRIGGER when: the developer wants to move, switch, or migrate mem0 off OSS/self-hosted to the hosted API, e.g. "migrate my mem0 setup to the platform", "switch…
forget
Remove a personal fact, preference, relationship detail, topic, or prior event from Muse's active memory and stop existing copies or automations from bringing it back. Use for explicit requests such as 'forget that', 'don't remember this about me', or 'remove that from your memory'. Do not use when 'forget it' merely…
Cortex
LifeOS memory, prior-session recall, and learning capture. USE WHEN capture this learning, document this learning, what did we learn recently, recent learnings, cortex, knowledge, knowledge base, search knowledge, what do we know about, archive, harvest, knowledge status, develop note, add to knowledge, add a people…
memory
Use when the user asks to remember, recall, forget, update, search, or inspect durable OpenSquilla memory, including profile facts in USER.md and long-term notes in MEMORY.md or memory//.md.