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 commands/pashki975/thm-claude-kit/notesgit clone --depth 1 https://github.com/pashki975/thm-claude-kitWrote 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/commands/pashki975/thm-claude-kit/notes)<a href="https://agentmods.dev/commands/pashki975/thm-claude-kit/notes"><img src="https://agentmods.dev/badge/commands/pashki975/thm-claude-kit/notes.svg" alt="Measured on agentmods" 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 | $0.00014 | $0.00308 |
| Opus 5 | $0.00007 | $0.00154 |
| Sonnet 5 | $0.00003 | $0.00062 |
| Haiku 4.5 | $0.00001 | $0.00031 |
Grade A, and why
notes 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 4d 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.
What it actually says
Update notes.md as the room's working memory (the "room brain"), not a passive log. Review scans/ and our conversation, then keep these sections current and factual — only what's confirmed.
# <Room> — Room Brain
## Goal / stop condition
- (flag string? user+root? specific answer?)
## Kill-chain position
- recon | foothold | lateral/pivot | privesc | domain/loot ← mark where we are
## Confirmed facts
- Target IP / hostname / OS:
- Open ports & services (with versions):
- Tech stack:
## Creds / keys held ← most important on complex rooms
| cred | type | works on | tried elsewhere? |
|------|------|----------|------------------|
## Hosts / network
- (every host/subnet found; note reachable vs pivot-only)
## Open leads
- (seen but not yet chased)
## Dead ends
- (tried, didn't pay off — don't loop back)
## Flags
- User:
- Root:
When called, reconcile this with everything we've learned: add new confirmed facts, record every cred and where it does/doesn't work, list new hosts and leads, cross off dead ends, and update the kill-chain position. This file is what lets the coach reason about the room as a whole rather than reacting to the last command.
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.
- 4d ago First seen · 44 lines · 14 tokens per session scan A 0a4fabd53ffe
notes is a command published in the GitHub repository pashki975/thm-claude-kit (2 stars, last pushed 21d ago), licensed MIT. It adds 14 tokens to every session and 308 once invoked, about $0.0001 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 commands, from other repositories
unwrap
You are resuming a Claude Code session. Execute all steps below in order — no confirmation needed.
wrap-find
Search session logs for a keyword or phrase. The search term is everything after /wrap-find in the user's message.
memory-gc
Inspect or rotate hunt-memory JSONL files (audit.jsonl, patterns.jsonl, journal.jsonl). Caps file size and keeps N rotated backups so memory does not grow unbounded.
remember
Log current finding or successful pattern to hunt memory. Auto-fills from /validate output if available. Usage: /remember.
context
Search memories by keyword and show relevant context.
sync
Two-way sync between Nemp Memory and CLAUDE.md.