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 cynthiajones34/GBrain --skill bos-cost-optimizationgit clone --depth 1 https://github.com/cynthiajones34/GBrainWrote 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/cynthiajones34/gbrain/bos-cost-optimization)<a href="https://agentmods.dev/skills/cynthiajones34/gbrain/bos-cost-optimization"><img src="https://agentmods.dev/badge/skills/cynthiajones34/gbrain/bos-cost-optimization/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/cynthiajones34/gbrain/bos-cost-optimization"><img src="https://agentmods.dev/badge/skills/cynthiajones34/gbrain/bos-cost-optimization.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 | $0.00031 | $0.02187 |
| Opus 5 | $0.00015 | $0.01094 |
| Sonnet 5 | $0.00006 | $0.00437 |
| Haiku 4.5 | $0.00003 | $0.00219 |
Grade A, and why
bos-cost-optimization 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 8d 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 — 182 lines — stays where its author put it; the contents beside it link to each section on GitHub.
BOS Cost Optimization Playbook
This skill is the operating manual for keeping the BOS gbrain deployment at floor cost. Every optimization in here has been tested on a single-operator scale (~500–5K pages, 1K–5K queries/mo). Read it before any change that touches search mode, embedding provider, or synthesis model.
The target: $0–5/mo for a single operator
The default gbrain install assumes enterprise scale (100K+ pages, 100K+ queries/mo, multi-user). For a single-operator personal brain, every default is over-provisioned. The minimum-cost posture is:
| Component | Minimum-cost choice | Monthly cost at 1K queries/mo |
|---|---|---|
| Database | PGLite embedded | $0 |
| Embeddings | Google Gemini free tier (gemini-embedding-001 at 768d) |
$0 |
| Reranker | (skipped — vector + BM25 + RRF only) | $0 |
| Search mode | conservative |
$0 (smaller chunks → fewer synthesis tokens) |
| Synthesis model | Haiku 4.5 | $0.80–1.50 |
| LLM expansion | OFF | $0 |
| Cron enrichment | Once weekly, not nightly | $0 |
| Backup storage | Local on the Railway volume | $0 |
Floor: ~$1/mo with active use. Ceiling: ~$5/mo at moderate use. Anything above $10/mo means a knob has drifted.
Why no reranker
The upstream gbrain default uses ZeroEntropy's bundled reranker. ZeroEntropy paused new signups as of August 2026, so this fork ships without one. The cost is small (~3-5% P@5 hit) and the floor is $0 lower.
If ZeroEntropy reopens: re-enable by setting search.reranker enabled=true in config and exporting ZEROENTROPY_API_KEY. Compare quality before/after with gbrain eval retrieval-quality and decide whether the reranker quality lift is worth the API key.
The 7 levers, ranked by $/quality impact
- Default to Google Gemini free tier.
gemini-embedding-001at 768d, free tier allows ~1,500 requests/day. That's a 500-page brain with normal use, and the daily limit means we won't hit it from a single import. Use your existing Google account — no new signup, no card, no email verification loop.
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.
- 8d ago First seen · 182 lines · 31 tokens per session scan A 53c137efdd3c
bos-cost-optimization is a skill published in the GitHub repository cynthiajones34/GBrain (0 stars, last pushed 1mo ago), licensed MIT. It adds 31 tokens to every session and 2,187 once invoked, about $0.0002 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
memory-proactive
Proactive layered recall and generic domain-aware routing.
memory-archivist
A set of scripts for archiving conversations, syncing them to a knowledge graph, updating summaries, and managing stored memories over time. A knowledge graph is a linked collection of information and relationships.
memory-starter-kit
Historical starter note for the memory sidecar stack.
mind
Local project memory with recall, provenance, policy, and dreams.
personal-knowledge-graph
Use when maintaining a LoomKG/Obsidian knowledge graph.
graph-mutation-plan
Cookbook for composing an applygraphmutations plan — stable entitykey patterns, the canonical label/edge vocabulary, evidence/invalidation/confidence discipline, and a worked example. Load this when building a non-trivial mutation plan.