Core behavioral rules and skills for NanoClaw personal assistant agents. Always-on rules for communication, verification, memory, and formatting.
97
97%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Advisory
Suggest reviewing before use
Core behavioral rules and skills for NanoClaw personal assistant agents. Always-on rules for communication, verification, memory, and formatting.
tessl install jbaruch/nanoclaw-core| Rule | Summary |
|---|---|
| core-behavior | Baseline conversational behavior — concise replies, no trailing summaries |
| telegram-protocol | Telegram-channel formatting and reply conventions |
| language-matching | Match the user's language; don't volunteer translations |
| default-silence | Silent-on-no-news as the default; surface only what's actionable |
| temporal-awareness | Convert relative dates, respect timezones, never schedule reminders for the past |
| ground-truth | Live source > cached/recalled state; always verify before acting |
| read-full-content | Read full email/message bodies, never just snippets/previews |
| context-recovery | Recover from compaction by reading state, not by claiming lost context. Includes post-compaction skill-block disambiguation (history vs new tasks) |
| post-compaction-trust | Compaction summaries are trusted state; do not re-derive what's already there |
| tone-matching | Mirror the user's register; tone shifts signal mode shifts |
| query-size-limits | Bounded SQL/API queries; never SELECT * on large tables |
| progress-updates | Brief updates at key moments; silence is not a failure |
| Skill | Description |
|---|---|
| status | One-line status emit on a defined cadence |
| check-unanswered | Two-phase scan for unanswered user messages — SQL filter + Phase 2 inline-answer reasoning |
alwaysApply: true. The agent follows these conventions without being prompted.nanoclaw-admin.ground-truth).See CHANGELOG.md for version history.