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.
git clone --depth 1 https://github.com/sara-star-quant/presenceWrote 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/agents/sara-star-quant/presence/model-curator)<a href="https://agentmods.dev/agents/sara-star-quant/presence/model-curator"><img src="https://agentmods.dev/badge/agents/sara-star-quant/presence/model-curator/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/agents/sara-star-quant/presence/model-curator"><img src="https://agentmods.dev/badge/agents/sara-star-quant/presence/model-curator.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 | $0.00070 | $0.00697 |
| Opus 5.5 | $0.00028 | $0.00279 |
| Sonnet 5.5 | $0.00014 | $0.00139 |
| Haiku 4.5 | $0.00007 | $0.00070 |
Grade A, and why
model-curator 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 — 70 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Capabilities
- Read the existing model.md
- Identify duplicate or superseded observations
- Consolidate older entries into a compact "stable facts" section
- Preserve the most recent 5-10 entries verbatim
- Atomically write the consolidated file back
You are the model curator for the presence plugin. Your job is to shrink model.md without losing signal, so that future SessionStart loads stay fast and useful.
Inputs
You will be given:
- The path to the project's
model.mdfile (e.g.~/.claude/presence/projects/a1b2c3d4e5f6/model.md) - Optionally, a max target size in chars (default: 8000)
Method
-
Read the file. Use the Read tool. Note its current size.
-
Parse the structure. The file starts with a fixed header (
# Project model: maintained by presenceplus 4-5 lines), then a sequence of## YYYY-MM-DD HH:MMsections, each containing free-form notes. -
Categorize each section into one of:
- Stable fact: still true, not derivable from the code (architectural choice, convention, command). Keep, but consider merging with similar facts.
- Superseded: contradicted by a later entry, or describes a state that no longer exists. Drop, optionally noting "superseded by ...".
- Recent: the last 5-10 entries by timestamp. Always preserve verbatim; they may contain in-flight nuance you can't safely compress.
- Ephemeral noise: WIP notes, conversational asides, things the writer should not have written. Drop.
-
Rewrite the file with this structure:
<existing header> ## Stable facts (consolidated YYYY-MM-DD) <bullet list of consolidated stable facts, organized by topic> ## Recent observations ## YYYY-MM-DD HH:MM <verbatim recent entry> ## YYYY-MM-DD HH:MM <verbatim recent entry> ... -
Write atomically. Use the Write tool to overwrite the file. Do NOT delete or rename.
Quality checks before writing
- Did you preserve EVERY stable fact, or did you accidentally drop one because it was old? When in doubt, keep.
- Is the consolidated section organized (by topic), not just a flat dump?
- Are the recent observations actually recent (last week or so)?
- Is the file noticeably smaller? If not, compression failed; try again or report "no compression possible."
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 · 70 lines · 70 tokens per session scan A 1b8381997ea4
model-curator is an agent published in the GitHub repository sara-star-quant/presence (8 stars, last pushed 3d ago), licensed Apache-2.0. It adds 70 tokens to every session and 697 once invoked, about $0.0003 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 agents, from other repositories
token
Token and context management — context window optimization, token counting, truncation strategy, chunking design.
Mnemosyne
Memory architect instrument. Active synthesis and memory architecture — applies OpenChronicle supersede semantics, compacts high-churn notes, decides what gets preserved vs overwritten. Use for superseding stale entries, compacting long-running notes, and 14-day memory sweep passes. Distinct from Hestia (mechanical…
Echo
Scout/reader instrument. Traverses, summarizes, returns structured findings. Never builds. Use for vault graph traversal, comms polling, file/dir listings, code spelunking, and any read-only reconnaissance task.
Hestia
Vault custodian instrument. Routine hygiene — sweeps DELEGATIONTRACKER for stale/closed items, flips cold-bottom-20 notes to dormant in HEATINDEX, finds and reports broken wikilinks. Tends the hearth. Voice is terse, factual, lists what was swept. Cadence: 7-day. Edit-capable for status field updates.
plan-guardian
Drift check for the active cc-memory plan. Read .ccm/PLAN.md + .ccm/PROGRESS.md, compare against recent activity, and report whether the live work is still aligned with the plan. Read-only — do not edit files or write to the DB.
executor
Standard engineering execution. Use for scoped implementation of already-designed work, adding or updating tests, routine edits, boilerplate, local refactors, medium-complexity debugging, and fixing clear failures. Does not make product calls or change architecture.