Files
bare-operating-system/packages/bare-os-seeder/kernel/share/agent-workspace/PROMPT.md
T

2.3 KiB

PROMPT.md - Reusable Command Templates

/remember

memory_append a FACT/CHECK/RISK/HANDOFF so the next session can find it.

/loop

schedule_task with interval 5m/2h/1d (or cron) and a prompt. unschedule_task to drop it.

/implement

Discover with glob/grep/read, track with todo_write, edit with unique search_replace or apply_patch, verify, then task_complete. NEVER ASK — just do it.

/grep

Use the grep tool (pattern + optional glob/path/context). Do not shell out to grep.

/patch

Emit one apply_patch document with *** Begin Patch / *** End Patch for multi-file edits.

/plan

enter_plan_mode, write ~/.agent/plan.md (blast radius, files, checks), exit_plan_mode, then implement unless the user asked for a plan only.

/fix

Reproduce with read_file / run_command / read_proc_file, apply the smallest unique edit, re-verify. Stop after three identical failures and report evidence.

/os-status

runtime_diagnostic_bundle plus get_swarm_peers. Summarize drive health, peers, kernel features, and recent logs.

/p2p-debug

Analyze current Hyperswarm swarm and suggest optimizations or issues.

/ops-diagnose

Inspect initd unit phases, timer drop-ins, cron/audit logs, boot policy, and kernel extension resolution; summarize root causes and next actions.

/reasoning-on

Enable process visibility with edit_agent_config (show_reasoning=true, reasoning_mode=trace) and keep output bounded.

/reasoning-off

Disable reasoning/process output with edit_agent_config (show_reasoning=false, reasoning_mode=off).

/xai-debug

Validate xAI compatibility settings, stream behavior, reasoning summary visibility, and tool-call handling. Recommend reasoning_mode=trace when summaries are absent.

/verify-hints

Pass verification_hints with the paths or topic you are about to edit (host checkout) to list suggested npm/node checks before leaving the agent.

/diag-bundle

Run runtime_diagnostic_bundle once for live /proc and resource snapshot before deep P2P or booter debugging.

Multi-agent handoff (human or tools)

Split work: one pass for repo/doc scout (read_skill + read_man_page), one for verification hints and smallest test lane, one for summarizing risks. Hand off structured JSON notes in MEMORY.md rather than repeating full logs.