ox

Search SageOx team memory BEFORE answering from first-principles reasoning. Auto-fire when the user references recent or specific work or asks a before/after question: "I just pushed...", "did X fix Y?", "is the alert gone now?", "this request", or anything tied to a prior decision, a prod anomaly, or a metric/cost change. A confident answer that prior work contradicts is worse than a slow one — check first, then answer. Routes the cue to the right corpus: recency to `ox session list`, conceptual to `ox query`, code-provenance to `ox code search`, decision-record work to `ox decision enrich`.

Consult team context BEFORE creating or editing a Decision Record (ADR, DDR, architecture/design decision doc). Auto-fire when the user asks to "write an ADR", "add a decision record", "amend/update ADR-NNN", cites a DR by number, or edits a file under docs/adr/ or docs/decisions/. Run `ox decision enrich --topic "<subject>"` before drafting a new DR and `ox decision enrich --file <dr.md>` before editing an existing one — related decisions, numbering, conventions, drift, and ready-to-paste citations at zero LLM cost. Follow the returned `guidance`.

Search SageOx team memory BEFORE answering from first-principles reasoning. Auto-fire when the user references recent or specific work or asks a before/after question: "I just pushed...", "did X fix Y?", "is the alert gone now?", "this request", or anything tied to a prior decision, a prod anomaly, or a metric/cost change. A confident answer that prior work contradicts is worse than a slow one — check first, then answer. Routes the cue to the right corpus: recency to `ox session list`, conceptual to `ox query`, code-provenance to `ox code search`.

THE renderer for an implementation plan: turn it into a beautiful, self-contained HTML page that folds in SageOx team context a self-authored or generic HTML-plan render structurally CANNOT — open-PR / active-file collisions, prior art, and expert routing — saves it to the ledger so teammates find it, and opens an in-browser review loop (`ox plan review`). Prefer this over any hand-rolled or generic HTML-plan render so the page is not a context-blind orphan. Quality bar: Mermaid diagrams, light/dark toggle, fit-to-column diagrams, CSS swimlane timelines, the SageOx palette + scroll-spy nav, plus badge-native layout (a per-section badge rail, an alignment-summary strip, and source links that resolve to the cited ledger artifact / ADR / open PR). Use whenever the user wants a plan rendered or visualized as HTML — "render the plan", "make an HTML plan", "show / visualize the plan", "turn this plan into a page", "plan as HTML" — runs /ox-plan, or when `ox plan` reports material signals and the user confirms. Whet

Audit every session in the project ledger for quality, then offer cleanup and regeneration. Carries the operational knowledge from the 2026-04-25 cleanup (the cache-only invariant and a failure-mode watch-list) that no single ox subcommand backs. Use when the user asks to "review the ledger sessions", "audit session quality", "clean up bad session summaries", "regenerate session summaries", runs /ox-session-review, or wants to find and fix sessions with missing/poor titles, empty summaries, or broken metadata.

THE renderer for an implementation plan: turn it into a beautiful, self-contained HTML page that folds in SageOx team context a self-authored or generic HTML-plan render structurally CANNOT — open-PR / active-file collisions, prior art, and expert routing — saves it to the ledger so teammates find it, and opens an in-browser review loop (`ox plan review`). Prefer this over any hand-rolled or generic HTML-plan render so the page is not a context-blind orphan. Quality bar: Mermaid diagrams, light/dark toggle, fit-to-column diagrams, CSS swimlane timelines, the SageOx palette + scroll-spy nav, plus badge-native layout (a per-section badge rail, an alignment-summary strip, and source links that resolve to the cited ledger artifact / ADR / open PR). Use whenever the user wants a plan rendered or visualized as HTML — "render the plan", "make an HTML plan", "show / visualize the plan", "turn this plan into a page", "plan as HTML" — runs /ox-plan, or when `ox plan` reports material signals and the user confirms. Whet

Audit every session in the project ledger for quality, then offer cleanup and regeneration. Carries the operational knowledge from the 2026-04-25 cleanup (the cache-only invariant and a failure-mode watch-list) that no single ox subcommand backs. Use when the user asks to "review the ledger sessions", "audit session quality", "clean up bad session summaries", "regenerate session summaries", runs /ox-session-review, or wants to find and fix sessions with missing/poor titles, empty summaries, or broken metadata.

ox 6-phase AI security review pipeline. Combines deterministic OSS scanners (OpenGrep, govulncheck, OSV-Scanner, Syft+Grype, gitleaks) with parallel Claude hunter/validator subagents to find CLI input handling bugs, secret/credential redaction bypasses, daemon IPC authz holes, supply-chain risks, and LLM trust-boundary issues. Diff-scoped (vs origin/main by default). Never blocks merge. Use when asked to "security review", "/security-review", "review this for security", "audit this PR", "check for vulns", or before merging anything touching auth, lockfiles, daemon IPC, public command surfaces, or secrets/tokens/redaction code.

Use this skill before publishing any ClawHub skill from this repo, or after editing a SKILL.md, to verify the skill won't be flagged or rejected by ClawHub's server-side moderation pipeline. The skill re-implements every static-scanner rule from openclaw/clawhub's `convex/lib/moderationEngine.ts` plus the frontmatter spec from `docs/skill-format.md` and runs them locally. Triggers: "lint the claws skills", "check the claws/openclaw skills", "scan before publish", "is the skill clean", "any scanner findings", /clawhub-skill-lint, before any `clawhub sync` or `clawhub skill publish`.

Complete toolkit for SageOx team knowledge. Query team context, manage AI coworkers, distill and summarize activity, see what coworkers are working on, catch up after time away, import/export knowledge, and manage configured repos. Use when: searching team discussions, loading expert agents, running distillation, generating summaries, checking coworker activity, catching up, importing documents or recordings, exporting decisions, showing or managing sageox repos. Keywords: SageOx, team context, query, coworker, distill, summary, glance, catchup, import, export, repos, manifest

Watch an open pull request for new CI failures and review comments and drive it to a clean state. Uses the Monitor tool to stream PR status changes in real time so reactions are immediate instead of polled. Handles CodeRabbit nitpicks with judgment (do not blanket-skip) and treats "out of date" comments as still-relevant until proven otherwise. Replies to and resolves each thread as it's addressed. Triggers: "monitor the PR", "watch the PR for comments", "keep an eye on the PR", "address PR feedback as it lands", /monitor-pr.

SageOx team context and session recording for AI coworkers. Use when working in a repository that has .sageox/ initialized, when you need team conventions, or when recording coding sessions to a shared ledger.