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/nisus74/humanise/initgit clone --depth 1 https://github.com/Nisus74/humaniseWhat 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 | $0.00000 | $0.00572 |
| Opus 5 | $0.00000 | $0.00286 |
| Sonnet 5 | $0.00000 | $0.00114 |
| Haiku 4.5 | $0.00000 | $0.00057 |
Grade A, and why
init 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 yesterday.
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 — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.
humanise: init
Set up the user's voice without making them complete a profile before seeing value.
Quick setup, about three minutes
- Resolve the private profile path. In a Claude Code marketplace plugin, use
${CLAUDE_PLUGIN_DATA}/profile; otherwise use the configured path orprofile/beside the skill. If it is missing, copyprofile.template/there. For a direct skill install, copyconfig.example.ymltoconfig.ymlbesideSKILL.md. - Ask the user to paste one short piece they wrote and like. A sent email, post, message or document paragraph is enough. Ask them to remove anything sensitive first.
- Ask what the sample was trying to achieve and who read it. Do not ask a long style questionnaire.
- If the user supplied text to improve, create up to three versions:
- close: minimal change;
- direct: point and ask moved forward;
- conversational: warmer and closer to speech.
- Ask which version is closest and what they would change. The choice and correction are stronger evidence than abstract style adjectives.
- Save the original sample as
profile/sample-<channel>-<slug>.mdusingSAMPLE_TEMPLATE.md. - Write the observed decisions to
profile/voice-decisions.mdwithprovisionalconfidence. Record any rejected moves inprofile/negative-examples.md. - Fill only the known fields in
profile/identity.mdandconfig.yml. Leave unknowns explicit.
At this point, produce a useful personalised draft. Do not block on a full fingerprint.
Deeper setup, offered after value is visible
When the user wants better coverage:
- Use
scripts/corpus-questionnaire.mdto collect five to ten real samples across the channels they actually use. - Write
profile/soul.mdfrom concrete convictions, not generic values. - Keep
profile/absolute-rules.mdto explicit non-negotiables. Do not install engine defaults as personal absolutes. - Run
scripts/generate-fingerprint.mdto buildprofile/voice-fingerprint.md. Label every pattern confirmed, supported or provisional and name its evidence. - Add relationship-specific differences to
profile/relationships.mdonly when direct evidence supports them.
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.
- yesterday First seen · 49 lines · 0 tokens per session scan A 282855896bc0
init is a command published in the GitHub repository Nisus74/humanise (1 stars, last pushed 5d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 572 tokens. 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
pm-review
Review local git diff with reasoning + snippets; cross-check; dispose findings into the rules library.
test-e2e
Write end-to-end tests for critical user journeys.
test-integration
Write integration tests for component boundaries.
newpkg
Generate new package with Nx.
grill-me
Run an interactive one-question-at-a-time oral drill for research defence or active-recall study, escalating around weak answers and ending with a study sheet. Use when the user asks to be grilled, quizzed, or prepared for a viva, job talk, seminar, or exam. Not for a written critique; use $devils-advocate or a review…
export
Design public API exports.