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 instructions/shiro-0x/hersona/claude-mdgit clone --depth 1 https://github.com/shiro-0x/hersonaWrote 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/instructions/shiro-0x/hersona/claude-md)<a href="https://agentmods.dev/instructions/shiro-0x/hersona/claude-md"><img src="https://agentmods.dev/badge/instructions/shiro-0x/hersona/claude-md.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 5.1 | $0.00906 | $0.00906 |
| Opus 5 | $0.00453 | $0.00453 |
| Sonnet 5 | $0.00181 | $0.00181 |
| Haiku 4.5 | $0.00091 | $0.00091 |
Grade A, and why
hersona CLAUDE.md 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 6d 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 — 75 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
Guidance for AI agents (Claude Code, etc.) working in this repository. The authoritative rules live in CONTRIBUTING.md; this file is the short, must-follow summary.
Update rule (keep docs in sync)
When you make a change that affects features or contracts, check the README in the same change and update it if it has drifted. Do not leave README updates "for later".
Check both README.md and README.ja.md (keep EN/JA in sync) whenever you
change any of:
- CLI subcommands / flags
- the
/hersonaskill's command syntax, modes, or behavior - the attribute schema (
schema/attribute.schema.json) or attribute count / categories — the single source of truth for counts istests/catalog_counts.py(TOTAL_PUBLIC_ATTRIBUTES/PUBLIC_CATEGORY_COUNTS); update it first, then update README.md / README.ja.md anddocs/REFERENCE.en.md/docs/REFERENCE.md(home of the full attribute catalog table) to match. CI'sscripts/check_readme_counts.pygate fails the build if any of those four files drift from it (added after an external review caught README/About/actual-count mismatches — do not hardcode the current number in this file either) - the public API (
hersona.core/docs/PUBLIC_API.md) - export formats / framework integrations
- any new user-facing file or doc (e.g. adding
REFERENCE.md→ link it from the README)
Keep the READMEs short. README.md / README.ja.md are the front page
(essentials only); detailed reference material (full attribute catalog, schema
fields, CLI walkthrough, skill recipes) belongs in docs/REFERENCE.en.md /
docs/REFERENCE.md — extend those instead of growing the README.
Then add a ## [Unreleased] entry in CHANGELOG.md and run
python scripts/validate.py and pytest.
If you add or remove attribute YAMLs, also run:
python scripts/build_site.py
python scripts/gen_checksums.py
build_site.py regenerates docs/app/data.json (the site data file).
gen_checksums.py regenerates checksums.json (the SHA-256 manifest
hersona update verifies downloads against — see docs/SECURITY.md). CI's
build_site.py --check and gen_checksums.py --check gates fail the build
if either file is stale. validate.py and pytest do not cover
these steps.
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.
- 6d ago First seen · 75 lines · 906 tokens per session scan A 064b7f4ab56c
hersona CLAUDE.md is an instructions file published in the GitHub repository shiro-0x/hersona (50 stars, last pushed 9d ago), licensed MIT. It adds 906 tokens to every session, about $0.0045 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 2026-08-30.
Other instructions, from other repositories
gleaner AGENTS.md
AGENTS.md instructions for liuqiaodongdong/gleaner, covering agent 使用约定（gleaner skill + cli）, 装机（仓库 ↔ skill，必须成对）, 根目录与 skill, 或 and 首次 / 采集前：必须做部署引导.
gogh AGENTS.md
Instructions for sirkon/gogh, covering gogh — llm usage context, what gogh is, mental model (read this first), minimal end-to-end example and entry points.
common-thread-mcp CLAUDE.md
Claude Code instructions for skyphusion-labs/common-thread-mcp, covering claude.md -- common-thread-mcp, what this is, documentation map, commands and auth model.
open-helplines CLAUDE.md
Claude Code instructions for Kouki-odaka/open-helplines, covering open-helplines — claude code instructions, project purpose, repository layout, non-negotiable rules (apply to all code changes) and common tasks.
openchat AGENTS.md
AGENTS.md instructions for opencoredev/openchat, covering project knowledge base, structure, where to look, code map and conventions.
openchat CLAUDE.md
Claude Code instructions for opencoredev/openchat, covering openchat - claude code instructions, project structure, routing, tech stack and authentication flow.