Every devrel verb.

Interactive wizard: scaffold .devrel/, configure LLM key, run health check, edit voice.md, generate first draft. --skip-chain, --skip-draft, --non-interactive available.

Pick provider (Anthropic / OpenRouter), validate with one-token ping, write .devrel/.env chmod 600. --rotate, --no-validate, --non-interactive.

Health check: state.db schema, LLM key validity, KB presence, voice/style/slop completeness. Names the fix verb on each failure.

Upgrade state.db schema after a devrel-origin version bump.

Read or write a single config.toml field without opening the file.

Execute the full weekly cycle: all 15 agents in dependency order, populate deliverables/<iso-week>/, write OKR rollup to state.db.

Register / inspect / remove the user-crontab entry that runs devrel run on cadence.

Per-run and rolling token spend across providers. Compares to the budget cap in config.toml.

Run Sage alone. GitHub issue priority + churn-risk scoring.

Run Echo alone. Reddit + HN + X mention harvest with batch sentiment.

Run Iris alone. Theme extraction from collected signals.

Run Nova alone. A/B test pre-registration with power analysis.

Run Rex alone. Competitive landscape + threat scoring.

Generate a single piece via Kai (8-stage editorial pipeline), or audit an existing draft against voice / style / slop.

Run Dex alone. AST-walks your code, generates API reference + architecture docs against your KB.

Run Vox alone. Generate screen-recorded walkthrough scripts + TTS narration.

Run Mox alone for any cross-channel campaign artifact.

Run Pax alone. Outreach sequences or competitor battle cards.

Run the growth pipeline: Nova hypotheses, Cyra funnel drop-offs, Argus post-publish recs.

Run Cyra alone. Pulls PostHog funnel time-series + drafts 3 ICE-scored A/B hypotheses.

Harvest a single URL via Firecrawl into .devrel/kb/, list what's there, or re-harvest configured sources. JS-rendered React docs may need manual markdown lifting (Firecrawl returns nav-chrome only on those).

PostHog + GitHub + Instantly rollups. Calibrated recommendations, week-over-week diffs.

Run Argus alone. Post-publish content performance + double-down / retire / rewrite recommendations.

Browse what shipped this week / past weeks. Each entry has its trace JSON.