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 skills add khou/gardenkit --skill garden-voicegit clone --depth 1 https://github.com/khou/gardenkitWrote 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/skills/khou/gardenkit/garden-voice)<a href="https://agentmods.dev/skills/khou/gardenkit/garden-voice"><img src="https://agentmods.dev/badge/skills/khou/gardenkit/garden-voice.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 | $0.00092 | $0.01298 |
| Opus 5 | $0.00046 | $0.00649 |
| Sonnet 5 | $0.00018 | $0.00260 |
| Haiku 4.5 | $0.00009 | $0.00130 |
Grade A, and why
garden-voice 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 7d 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 — 138 lines — stays where its author put it; the contents beside it link to each section on GitHub.
garden-voice
Builds and maintains ~/garden/meta/voice.md: the user's voice profile derived from their own messages. Loaded on-demand by drafting tasks for tone-matching.
Why this exists
Asking a user to describe their writing voice in the abstract gives shallow results. Sampling their actual messages and synthesizing patterns gives the agent a concrete, calibrated reference. Per the source video, having a voice profile is reportedly the single biggest output-quality multiplier for any drafting work.
Modes
init
First-time bootstrap. Sample broadly across the user's recent sent messages.
- Find user ID via
slack_search_users(use the user's name or email frommeta/user.md). - Pull sent messages. Use
slack_search_public_and_privatewith queryfrom:<USER_ID>. Paginate to gather ~500–1000 messages. Skip threads where the only content is reactions or single-emoji replies. - Aggregate text into a single corpus.
- Redact (see redaction rules below): strip proper nouns, numbers, URLs, but preserve all linguistic content.
- Synthesize patterns: observe and write down:
- Sentence rhythm (terse / flowing / fragments)
- Capitalization habits (lowercase casual / sentence case / mixed)
- Punctuation tics (em-dashes, ellipses, parenthetical asides, trailing
?for soft asks) - Vocabulary preferences (favored words, slang, technical register)
- Emoji and reaction patterns (which, how often, in what positions)
- Greeting / closing patterns
- Sentence starters and connector words
- Tone markers (irony, hedging, directness)
- Pick 10–15 anchored examples: short snippets (≤2 sentences each, redacted) that demonstrate the patterns concretely.
- Write
~/garden/meta/voice.mdper the output format below. - Commit to the vault git repo with message
voice: init from <N> Slack messages.
refresh
Periodic update: appends new observations rather than overwriting.
- Read existing
voice.mdto know what patterns are already noted. - Pull last ~200 messages since the file's
updated:date. - Diff: any new patterns or notable shifts in style?
- Update
voice.md: refresh anchored examples (drop oldest, add freshest), append new pattern observations. - Commit with
voice: refresh: <summary of changes>.
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.
- 7d ago First seen · 138 lines · 0 tokens per session scan A e791bec026ce
garden-voice is a skill published in the GitHub repository khou/gardenkit (4 stars, last pushed 3mo ago), licensed MIT. It adds 92 tokens to every session and 1,298 once invoked, about $0.0005 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 skills, from other repositories
ara-research-manager
Records research provenance as a post-task epilogue, scanning conversation history at the end of a coding or research session to extract decisions, experiments, dead ends, claims, heuristics, and pivots, and writing them into the ara/ directory with user-vs-AI provenance tags. Use as a session epilogue — never during…
loop-engineering
Use when a repeatable task must become a bounded Trigger -> Execute -> Verify -> State loop, scheduled automation, goal agent, or metric-driven research cycle.
harness-engineering
Use when an agent workflow needs production-like runtime controls for context, tools, permissions, observability, scheduling, evaluation, recovery, or maintenance.
agent-teams-command
Use when work has genuinely independent streams or distinct builder, evaluator, domain, and integration roles that require bounded multi-agent command scaled from 5 to 100+ agents.
agentic-engineering
Use when designing or refactoring a model-native engineering workflow with bounded autonomy, probes, custom evaluation, durable state, and verified write-back.
anthropic-os
Use when a personal or team operating system needs a bounded redesign using Four-C, closed-loop controls, 70/30 allocation, 3B creativity, experiments, and prediction-error learning.