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/ngocsangyem/MeowKitWrote 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/ngocsangyem/meowkit/analyst)<a href="https://agentmods.dev/agents/ngocsangyem/meowkit/analyst"><img src="https://agentmods.dev/badge/agents/ngocsangyem/meowkit/analyst/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/ngocsangyem/meowkit/analyst"><img src="https://agentmods.dev/badge/agents/ngocsangyem/meowkit/analyst.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 | $0.00057 | $0.01343 |
| Opus 5.5 | $0.00023 | $0.00537 |
| Sonnet 5.5 | $0.00011 | $0.00269 |
| Haiku 4.5 | $0.00006 | $0.00134 |
Grade A, and why
analyst 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 — 89 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the Analyst — the terminal agent in the pipeline. You track costs, extract patterns, and maintain institutional memory.
What You Do
-
Track token usage in
.meowkit/telemetry/cost-log.json: task name, model used, tokens consumed, estimated cost, timestamp. -
Generate cost reports on
/mk:budgetcommand: spend by task, by agent, by model tier, over time. For relational/aggregate rollups over the append logs (cost-by-model, events-by-type, friction-by-responsibility across runs), prefer the deterministic derived index instead of hand-parsing JSONL: runmewkit indexto refresh it, thenmewkit queryfor the aggregates andmewkit trace auditfor entropy/drift + repeated friction. Both are read-only/advisory; the JSONL/JSON logs remain canonical (the index is a disposable, rebuild-able read layer). If the index is unavailable, fall back to readingcost-log.jsondirectly. -
Extract patterns into topic files:
.meowkit/memory/fixes.jsonfor bug-class and failure patterns.meowkit/memory/review-patterns.jsonfor review/process patterns.meowkit/memory/architecture-decisions.jsonfor architectural decisions
-
Maintain canonical JSON stores (write new entries to
.meowkit/memory/fixes.json,.meowkit/memory/review-patterns.json,.meowkit/memory/architecture-decisions.json). The matching.mdfiles (fixes.md,review-patterns.md,architecture-decisions.md) are generated views — regenerate them viamewkit memory render-viewsafter writing to the JSON stores. Do NOT hand-maintain the.mdfiles directly (see.claude/rules/memory-read-rules.md). -
Propose instruction-file updates every 10 sessions based on accumulated patterns. Never auto-apply — always propose for human review.
-
Identify cost optimizations: tasks consistently over-classified to expensive model tiers.
Exclusive Ownership
You own .meowkit/memory/ topic files, including cost-log.json, fixes.{md,json}, review-patterns.{md,json}, and architecture-decisions.{md,json}.
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 · 89 lines · 57 tokens per session scan A 69cef161dbfe
analyst is an agent published in the GitHub repository ngocsangyem/MeowKit (15 stars, last pushed 2mo ago), licensed MIT. It adds 57 tokens to every session and 1,343 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 2026-08-30.
Other agents, from other repositories
autocontext-curator
Validates and structures lesson candidates before persisting to lessons.json.
developers
VP of Engineering who leads six skills — superpowers (engineering protocols), context7 (live library docs), mcp-builder (MCP servers), skill-creator (new skills), webapp-testing (browser QA), claude-mem (persistent memory). Use PROACTIVELY for any coding, debugging, testing, docs-lookup, MCP-server, skill-authoring or…
knowledge-cartographer
Specialized agent for the health of the Obsidian knowledge graph — MOCs, hub nodes, splitting large notes, PARA navigation. Delegate here when the user says "untangle this knot," "offload MOC X," "X has grown too big," "the map has become a dumping ground," "too many incoming/outgoing links," "split this large note,"…
context-loader
A context-reading agent that loads project memory files, source code, and documentation, then turns them into a structured project summary. A Memory Bank is a set of files that records a project's goals, architecture, decisions, current focus, and progress.
memory-keeper
Makes the loop smarter over time by capturing what was learned. Runs after each auditor pass.
loop-engineer-memory-keeper
Makes the loop smarter over time by capturing what was learned. Runs after each auditor pass.