Hone Administrator
Skill748 starsChanged 3 months agoMITView on GitHub
Admin-only skill for viewing or modifying Hone source code and configuration, and for performing restarts. Available only to users listed in config.yaml admins
We split this skill into the rules it gives an agent
10 rules. About 0 are common, 4 are distinctive, and 6 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 55
Read the target file first and confirm its current content
- Distinctive: line 58
If Rust source changed, restart the service so the change can take effect
- Very specific: line 59
If only config.yaml or skills/ changed, restart is still required because Hone loads them at startup
- Distinctive: line 65
When to use: after modifying Rust source or config so the change can take effect.
- Very specific: line 76
After the tool returns success=true, tell the user:
- Very specific: line 92
Read data/runtime/current.pid to see the process PID
- Very specific: line 93
Read data/logs/restart.log to see recent restart logs
- Very specific: line 94
Read data/logs/hone.log to see the runtime logs
- Very specific: line 102
Do not delete critical files: never delete config.yaml, Cargo.toml, or the data/ directory
- Distinctive: line 104
Admins only: if the user is not in the admin list, refuse all admin actions politely
≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Hone Administrator Skill (hone_admin)
Security note: This skill is available only to administrators listed in
config.yamlunderadmins. If an unauthorized user tries to trigger admin actions, politely refuse.
Project Structure Overview
<PROJECT_ROOT>/ <- project root (current working directory)
├── config.yaml <- main config file (LLM, channels, admin list, and so on)
├── bins/hone-cli/ <- canonical CLI startup path (`hone-cli start --build`)
├── bins/
│ ├── hone-imessage/ <- iMessage channel entrypoint
│ ├── hone-discord/ <- Discord channel entrypoint
│ └── hone-telegram/ <- Telegram channel entrypoint
├── crates/
│ ├── hone-core/ <- core config and Agent interfaces
│ ├── hone-channels/ <- shared channel logic (`HoneBotCore`)
│ ├── hone-tools/ <- tool implementations, including `restart_hone`
│ ├── hone-llm/ <- LLM call layer
│ └── hone-memory/ <- session persistence
├── agents/
│ ├── gemini_cli/ <- Gemini CLI agent
│ └── codex_cli/ <- Codex CLI agent
├── skills/ <- built-in system skills (`.md` files)
│ └── hone_admin/SKILL.md <- this file
├── data/
│ ├── runtime/current.pid <- PID of the current `hone-cli start` supervisor
│ └── logs/restart.log <- restart log
└── Cargo.toml <- Rust workspace configOperation 1: View or Modify Source Code or Config
Use the gemini_cli agent (default):
Gemini CLI runs in --yolo mode and can read and write files directly. Tell the user what you are about to do, then operate on the target files.
Common edit cases:
- Modify
config.yaml: adjust the LLM model, timeout, prompts, admin list, and so on - Modify
skills/<name>/SKILL.md: update a skill's instructions or prompt - Modify Rust source (
bins/,crates/): a rebuild is required before the change takes effect
Edit flow:
- Read the target file first and confirm its current content
- Make the change
- Show the diff or a summary to the user and confirm it looks correct
- If Rust source changed, restart the service so the change can take effect
- If only
config.yamlorskills/changed, restart is still required because Hone loads them at startup
Operation 2: Restart Hone
When to use: after modifying Rust source or config so the change can take effect.
Restart tool: restart_hone(confirm="yes")
Flow:
- Confirm the edits are done
- Tell the user: "Hone will restart shortly and be unavailable for about 1-3 minutes, including rebuild time"
- Call the tool:
restart_hone(confirm="yes") - After the tool returns
success=true, tell the user:- "Hone will stop and restart in about 3 seconds"
- "The restart includes a rebuild and is expected to take about 1-3 minutes"
- "The service will recover automatically after the restart; if it is still unresponsive after 5 minutes, check the machine manually"
Restart mechanism:
- The tool reads
data/runtime/current.pidto get the currenthone-cli startPID - It waits 3 seconds in the background so the reply can be sent first, then kills the process
- It immediately runs the source CLI build-and-start path in the project root
- The new PID is written to
data/runtime/current.pid
Operation 3: Check Runtime Status
You can use the following to check the current Hone state:
- Read
data/runtime/current.pidto see the process PID - Read
data/logs/restart.logto see recent restart logs - Read
data/logs/hone.logto see the runtime logs
Strict Rules
- Confirm before changing: show a diff or summary before restarting
- Restart is risky: during restart, iMessage and Discord replies may be unavailable
- Do not delete critical files: never delete
config.yaml,Cargo.toml, or thedata/directory - You must use the tool: restarts must go through
restart_hone; do not kill the process directly - Admins only: if the user is not in the admin list, refuse all admin actions politely
In this repository37 other files it gives its agents.
- AGENTS.mdAGENTS.md
- memory/agents.mdAGENTS.md
- alipay-aipaySkill
- event-engine-baseline-testingSkill
- event-engine-push-reviewSkill
- Formal ReleaseSkill
- Analyst CoverageSkill
- Chart VisualizationSkill
- Company Latest DevelopmentsSkill
- Company PortraitSkill
- company-thesis-ratingsSkill
- Deep Stock ResearchSkill
- Earnings ReadoutSkill
- earnings-researchSkill
- ETF AnalysisSkill
- First PrinciplesSkill
- FundamentalsSkill
- Gold AnalysisSkill
- hari-investSkill
- Image GenerationSkill
- Image UnderstandingSkill
- Industry MapSkill
- Influencer ViewsSkill
- Market AnalysisSkill
- Missed Digest ItemsSkill
- Moat AnalysisSkill
- Notification PreferencesSkill
- Options AnalysisSkill
- PDF UnderstandingSkill
- Portfolio ManagementSkill
- Position AdviceSkill
- Scarcity DifferentiationSkill
- Scheduled Task ManagementSkill
- Sector To StockSkill
- Skill ManagementSkill
- Stock ResearchSkill
- Valuation AuditSkill
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

