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 agentmods add skills/impactbrussels/founderos/capture-learningnpx skills add impactbrussels/FounderOS --skill capture-learninggit clone --depth 1 https://github.com/impactbrussels/FounderOSWhat 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 | $0.00082 | $0.00936 |
| Opus 5 | $0.00041 | $0.00468 |
| Sonnet 5 | $0.00016 | $0.00187 |
| Haiku 4.5 | $0.00008 | $0.00094 |
Grade A, and why
capture-learning 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.
How it starts
The opening of the file, as written. The whole thing — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Capture Learning - how the OS learns as it goes
Founder OS is meant to improve every time it's used. This skill is the mechanism: it turns a
real outcome into a durable, sourced lesson in knowledge-base/, so future sessions inherit
it instead of relearning it the hard way. This mirrors the discipline a good operator keeps -
write down what you learned the moment you learned it.
When to capture (and when not to)
Capture when a lesson is durable and transferable - true beyond this one session and useful to another founder. Do NOT capture one-off trivia, secrets/PII, or anything the founder hasn't confirmed actually happened. A draft outcome is not a lesson; a real result is.
The method
- Confirm the outcome is real. What concretely happened, and what's the evidence? (A test result, a customer reply, a metric move, a failed launch.) If it's speculation, don't record it.
- Distil to one transferable lesson. One sentence a stranger could apply. Strip the sector specifics - keep the principle (theme-agnostic).
- Classify it into one of the three KB files:
knowledge-base/PRINCIPLES.md- a durable truth about how startups/founders work.knowledge-base/DOS-AND-DONTS.md- a concrete do/don't for using Founder OS or AI well.knowledge-base/PLAYBOOK-LESSONS.md- a dated, situational lesson from real use.
- Write the entry in the file's format (below), with today's date and the source/evidence.
- Link it to the skill or flow it came from, so the lesson finds its way back into context.
Entry formats
PLAYBOOK-LESSONS.md (append under the right stage heading):
### YYYY-MM-DD - <one-line lesson>
- **Context:** <stage / what we were doing / which skill>
- **What happened:** <the real outcome + evidence>
- **Lesson:** <the transferable takeaway, one sentence>
- **So next time:** <the changed behaviour>
DOS-AND-DONTS.md (append a bullet under DO or DON'T):
- **DO/DON'T:** <imperative> - <why, one clause> (added YYYY-MM-DD)
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.
- 3d ago First seen · 80 lines · 82 tokens per session scan A 1287e0f50743
capture-learning is a skill published in the GitHub repository impactbrussels/FounderOS (2 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 82 tokens to every session and 936 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 2026-08-31.
Other skills, from other repositories
metrics-instrumentation
Specification for instrumenting an opik-backend workflow with operational OpenTelemetry metrics — per-stage throughput/latency/error counters and native histograms, dimensioned per-customer (workspace). Use when a pipeline (scoring, ingestion, experiments, jobs) needs per-stage visibility. Covers metric emission only…
review-agents-md
Audit Dograh AGENTS.md files for drift against the live repo and for bad scope boundaries between parent and child docs. Use when the user asks to review existing AGENTS files, identify stale guidance, decide whether a subtree needs its own AGENTS.md, or update the AGENTS.md hierarchy under the repo root, api/, or ui/.
twitter-hand-skill
Expert knowledge for AI Twitter/X management — API v2 reference, content strategy, engagement playbook, safety, and performance tracking.
chat-complex-documents
Chat with and search your complex documents — ask questions, extract tables and fields, and get answers grounded in the source. Connects the hosted Unstructured Transform MCP server to parse, structure, and enrich PDFs, Word/Excel/PowerPoint, images, scanned files, emails, and 60+ other formats into clean, AI-ready…
experiment-tracking-swanlab
Provides guidance for experiment tracking with SwanLab. Use when you need open-source run tracking, local or self-hosted dashboards, and lightweight media logging for ML workflows.
prompt-engineer
Prompt engineering expert for chain-of-thought, few-shot learning, evaluation, and LLM optimization.