メニュー
Check provider usage via TUI worker (~30s, costs CLI tokens). FALLBACK when `check-usage-http` returns RATE_LIMIT or fails. Spawns an ephemeral tmux session, runs `/usage` (or `/status` for Codex), parses the modal, kills the session.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Extract the company logo for a company in the companies table and store it as a small base64 data-URI (max ~35KB, min 32px). Primary path is fully automated via logo_fetch.py against the official website (apple-touch-icon → icon → og:image → favicon); when the site blocks bots or has no usable icon, find a direct logo image URL via web search and pass it with --from-url. Verify the website really belongs to the company BEFORE fetching. Sets companies.logo, logo_source, logo_fetched.
Lo sweep di manutenzione INFRA del Mantenitore 👷♂️ (gemello del Dottore, scope infrastruttura non agenti). Una passata giornaliera one-shot: canary di liveness dei processi salva-vita del container (bridge/daemon/watchdog) via process_health.py, smoke-test dei tool mission-critical (browser/LinkedIn) via tool_health.py, audit/consolidamento deps fuori standard, GC di script e tmp orfani, de-dup di script ricorrenti, freschezza deps, trend disco/RAM. Single-writer: il Mantenitore è l'UNICO che ripara l'infra; le azioni DISTRUTTIVE (delete/archive) le PROPONE, il Capitano decide. Esito in append su mantenitore-logbook.jsonl.
Doctor-only. Context-refresh round: for each agent session read its real context occupancy (provider client-side command, zero tokens) and refresh ONLY sessions whose context window is >50% full — do a retrospective (capture + interview + analytics), append a dense synthesis to the growing daily journal, then KILL + recreate + resume the session with continuation context, so its context window is cleared without losing where it was. Runs 2× per work window (at +30min and at mid). Skips fresh, low-context (≤50%), and Capitano-parked sessions.
Run the mandatory 3-round CV review loop with the Critico — autonomously, without going through the Capitano. For each round you spawn a FRESH `CRITICO-S<N>` session (same N as your Scrittore session: SCRITTORE-2 → CRITICO-S2), send PDF + JD, wait for the structured verdict, kill the Critic, correct the CV, regenerate the PDF, and start the next round with another fresh instance. Three rounds are non-negotiable — neither 1 nor 2. After the 3rd round, gate: `critic_score ≥ 5` → `ready`, else `excluded`. Owned by the Scrittore.
DB + filesystem contract every Scrittore follows when taking a position from `scored` (≥50) to `ready`/`excluded`. Three gates BEFORE writing a single line of CV (anti-rewriting, anti-collision, link verification), one canonical path for deliverables, one final gate after the 3rd Critic round. Skipping any of these produces duplicate work, overwrites another Writer's claim, or — worst — pushes an `excluded`-grade CV to the user as `ready`. Owned by the Scrittore.
Insert NEW records into the JHT DB (positions / scores / applications / companies / position_highlights). Use it ONLY when an agent needs to create a record — Scout for positions, Analyst for companies and highlights, Scorer for scores, Writer for applications. Never blind-overwrite — for updates use `db-update`.
| name | check-usage-tui |
| description | Check provider usage via TUI worker (~30s, costs CLI tokens). FALLBACK when `check-usage-http` returns RATE_LIMIT or fails. Spawns an ephemeral tmux session, runs `/usage` (or `/status` for Codex), parses the modal, kills the session. |
| allowed-tools | Bash(tmux *), Bash(python3 *) |
Solo come L2 dopo che L1 (HTTP) ha fallito. È robusta perché bypassa il rate-limit Anthropic (legge dal CLI locale), ma costa:
Ordine di fallback: HTTP → TUI → FATAL.
python3 -c "import json; print(json.load(open('/jht_home/jht.config.json'))['active_provider'])"
Nome unico col timestamp (no collisione):
SESS="SENTINELLA-WORKER-$(date +%s)"
tmux new-session -d -x 220 -y 50 -s "$SESS" -c /jht_home
tmux send-keys -t "$SESS" "export HOME='/jht_home'" C-m
tmux send-keys -t "$SESS" "export PATH='/app/agents/_tools:/jht_home/.npm-global/bin:\$PATH'" C-m
tmux send-keys -t "$SESS" "export KIMI_CLI_NO_AUTO_UPDATE=1" C-m
| Provider | Comando |
|---|---|
claude / anthropic | claude --dangerously-skip-permissions |
kimi / moonshot | kimi --yolo |
codex / openai | codex --yolo |
tmux send-keys -t "$SESS" "<comando>" C-m
sleep 18 # boot + eventuale trust dialog auto-accept
| Provider | Slash command |
|---|---|
claude / anthropic | /usage |
kimi / moonshot | /usage |
codex / openai | /status (NON /usage!) |
tmux send-keys -t "$SESS" "<slash>" Enter
sleep 4 # render modal
⚠️ NIENTE Esc preventivo prima del slash command: rompe Kimi (interpreta /usage come testo) e per gli altri è inutile in sessione fresh.
tmux capture-pane -t "$SESS" -p -S -100
Leggi il pane coi tuoi occhi LLM, estrai usage, reset, weekly.
Current session
████████ 45% used
Resets 8:10pm (UTC)
Current week (all models)
████ 25% used
Resets Apr 27, 5am (UTC)
→ usage=45, reset=20:10 UTC, weekly=25
╭─── API Usage ──────────────────────────────────╮
│ Weekly limit ━━ 60% left (resets in 2d 27m) │
│ 5h limit ━━ 100% left (resets in 3h 27m) │
╰────────────────────────────────────────────────╯
→ usage=100-100=0, weekly=100-60=40
→ reset_at: ora_UTC + 3h 27m → calcola HH:MM UTC
5h limit: [████░░] 35% left (resets 18:10)
Weekly limit: [██░░░░] 88% left (resets 14:00 on 02 May)
→ usage=100-35=65, weekly=100-88=12, reset=18:10
⚠️ Codex riporta "% left", non % used: sottrai da 100.
tmux kill-session -t "$SESS"
⚠️ Killa SEMPRE anche se il parse fallisce. Niente sessioni persistenti.
python3 /app/shared/skills/usage_record.py --manual \
--usage <X> --reset-at <HH:MM> --weekly <W> \
--provider <claude|kimi|openai> --source sentinella-worker
Hai ora il sample fresco con proj/status calcolati da compute_metrics. Aggiorna memoria (vedi skill memory-state) e decidi ordine al Capitano (vedi skill order-formats).
Enter (NON Escape: cancella e ti butta in bash).Loading usage data… infinito (Claude)TUI bloccato. Kill+respawn la sessione e riprova; se persiste cadi su L3 (FATAL).
Unrecognized command (codex)Hai mandato /usage su codex. Manda /status e basta.
Kill la sessione e respawna.