add-bib-entry[Codex | Claude Code]: Use when adding works to my bibliography, creating BibTeX/BibLaTeX entries, adding DOI/ISBN/URL references, resolving missing citekeys, replicating zotra-extras-add-entry followed by ebib-extras-process-entry, or preparing notes that cite works not yet in the configured bibliography files.add-to-emacs-packages[Codex | Claude Code]: Add or register one of the user’s Emacs packages in my-emacs-packages.org, create its documentation note, and add its GitHub profile README card. Use when the user wants to update the public Emacs packages list for a specific package.audit-mac-app[Codex | Claude Code]: Audit a macOS .app before running it or granting permissions, especially when the user asks whether a Mac app is safe, suspicious, malware, from an unknown developer, or requests screen recording, accessibility, camera, microphone, input monitoring, or other sensitive permissions.automate[Codex | Claude Code]: Design and build a reusable AI automation. Use for automate, create automation, turn this into a skill/command, new workflow, AI helper, or deciding whether a workflow should become a skill, directive, tool, or Emacs command.build[Codex | Claude Code]: Spec-based build workflow. Use for /build, build from spec, interview me about the spec, turn this rough spec into a plan, or developing spec.md into an implementation-ready plan before implementation.chrome-permission-audit[Codex | Claude Code]: Audit and revoke the sites Claude in Chrome is permanently allowed to access, across every Chrome profile. Use when the user asks to check/audit/review Claude’s browser site permissions, asks which sites Claude can access, wants to revoke browser access to sensitive sites, or asks whether the “permanent but not sensitive” policy still holds. Not for Chrome’s own content settings, extension host permissions, or OAuth app grants.code-audit[Codex | Claude Code]: Audit code for actual defects, including bugs, application security issues, error handling gaps, and correctness problems. Not for readability, architecture, test coverage, performance-only, or broad environment security reviews.config-audit[Codex | Claude Code]: Audit agent configuration for redundancy, conflicts, dead weight, staleness, and instruction drift across CLAUDE.md or AGENTS.md, skills, memory, hooks, settings, and docs.design-audit[Codex | Claude Code]: Audit code architecture and maintainability for unnecessary complexity, premature or missing abstractions, duplication, unclear boundaries, and refactoring opportunities. Not for bug, security, or readability-only reviews.diagnose[Codex | Claude Code]: Diagnose a specific agent, tooling, configuration, or workflow incident before any remedy. Use for /diagnose, why-did-this-happen questions, recurring process failures, broken hooks/skills/scripts, or diagnose-and-fix requests; not for ordinary code bugs, broad configuration audits, or repository-wide AI developer-experience reviews.document-elisp-extras[Codex | Claude Code]: Audit, create, or refresh documentation for all Emacs extras packages in the dotfiles repo. Use when the user asks to document every extras package, fill missing extras docs, refresh outdated extras docs, check extras documentation coverage, or bring emacs/extras/doc up to date with emacs/extras sources.document-elisp-package[Codex | Claude Code]: Create or update an Org manual for an Emacs Lisp package in the Denote manual style. Use for documenting Elisp packages or emacs/extras; use generate-readme for Markdown README generation alone.dotfiles-context[Codex | Claude Code]: Route changes to emacs/config.org, dotfiles extras, Elpaca-managed package checkouts, and paired Claude/Codex configuration. Use to prepare external-package fixes, choose the canonical edit location, tangle the active Emacs profile, or run paired-config checks; not for read-only or unrelated dotfiles work.download-paper[Codex | Claude Code]: Download the PDF of an academic paper (by DOI, publisher/doi.org URL, arXiv id, Anna’s Archive md5, or title) with `paper-fetch`, the single paper-acquisition tool, including its browser fallback for challenge-protected hosts. Use whenever the user wants a paper, article, preprint or its PDF obtained, fetched, downloaded, saved or attached, even if they only paste a DOI or link, mention Anna’s Archive, Sci-Hub, LibGen, PhilPapers or Unpaywall, or ask why a paper download failed. Not for books by ISBN or for creating bibliography entries alone (use add-bib-entry, which calls this for attachments).dx-audit[Codex | Claude Code]: Audit a codebase’s AI developer experience, including project instructions, verification commands, automation, onboarding traps, and collaboration friction. Not for ordinary bug, security, architecture, readability, or PR reviews.elisp-conventions[Codex | Claude Code]: Apply local coding and verification rules when modifying or testing Emacs Lisp, or when explicitly reviewing code against these conventions. Use for dotfiles extras, standalone Elpaca packages, and other .el files; not for non-Elisp work or read-only bug, security, or design reviews where these conventions are not material.emacs-freeze[Codex | Claude Code]: Diagnose a frozen, hung, beachballing, or unresponsive Emacs by sampling the process, reading server buffers when possible, and recommending recovery. Use when Emacs is stuck or not responding; do not use for ordinary Elisp errors while Emacs is responsive.end-to-end[Codex | Claude Code]: Use when the agent must execute or directly observe decisive live software acceptance through a real user-visible or runtime surface, including browser, Emacs, terminal display, async/network delivery, persistence, or scheduler behavior. Automation may drive the actual surface. Do not use solely to give manual test instructions; explain, write, or review; run an automated E2E suite; deploy or check deployment health/status; reproduce research results; or verify pure logic.find-skills[Codex | Claude Code]: Discover, vet, and optionally install agent skills when the user asks “find a skill for X”, “is there a skill that can…”, “how do I add support for X”, or otherwise wants reusable/installable agent capabilities. Do not use for ordinary task execution unless the user is explicitly exploring skills, tools, or capability extensions.fix-drive-errors[Codex | Claude Code]: Use when Google Drive reports sync errors, a symlink or generated directory exists under ~/My Drive, or a Google Docs pointer fails to upload from the synced personal-account root.generate-readme[Codex | Claude Code]: Generate or update a GitHub-facing README.md for an Emacs Lisp package from its Org manual. Use when the user says “generate readme”, “create readme”, “update readme”, “readme from manual”, or wants a concise Markdown README for an Elisp package. Do not use for general project README writing or for packages without an Org manual; use document-elisp-package first when the manual is missing.gitguardian-triage[Codex | Claude Code]: Triage and remediate GitGuardian secret incidents. Use for gitguardian/GG triage, secret audit, rotating leaked secrets, processing incidents, or clearing the GitGuardian dashboard.google-sheets-comments[Codex | Claude Code]: Handle Google Sheets comment review, replies, resolves, reopens, and notification emails correctly. Use this whenever the user asks to process, review, reply to, resolve, reopen, audit, verify, or summarize Google Sheets comments; mentions Google Sheets comment notification emails; asks to handle spreadsheet feedback; or opens/inspects a Google Sheet specifically to deal with comments. Do not use for ordinary spreadsheet editing or non-comment data analysis. This skill prevents treating Gmail notifications as the comment source of truth.handoff[Codex | Claude Code]: Use when the user asks to save next steps, prepare a resume prompt, or continue work in a fresh session of the current tool, or when clear follow-up work should resume there.humanize[Codex | Claude Code]: Rewrite English prose to remove formulaic AI-style wording while preserving meaning, voice, quotations, citations and markup. Use for /humanize, de-slop, removing AI tells, “sounds like ChatGPT/Claude/Gemini”, or final cleanup of AI-assisted prose. For writing under my name, use personalize, which incorporates this catalogue.install-mcp-server[Codex | Claude Code]: Install and configure a new MCP server for Claude Code, including discovery, auth verification, credential-safe setup, scope placement, registration, sync, and restart/testing expectations.interpretability-audit[Codex | Claude Code]: Audit code for intent transparency and reader comprehension. Use for understandable, readable, self-explanatory, naming, or future-maintainer clarity reviews; not for bug/security, architecture, DX, or lint-only reviews.lint-elisp[Codex | Claude Code]: Fix byte-compile warnings, checkdoc notes, and other Elisp diagnostics in one or more files. Use when the user says “lint elisp”, “fix warnings”, “fix diagnostics”, “elisp lint”, “checkdoc”, “byte-compile warnings”, or wants to clean up Elisp compiler/linter output.migrate-profile[Codex | Claude Code]: Run post-Emacs-profile-switch housekeeping after elpaca profile changes. Use when the user says “/migrate-profile”, changes Emacs or elpaca profiles, switches to a new profile, or wants Claude/Codex sessions, memory, trust entries, history, cwd paths, profile symlinks/caches, or new-profile repos checked after a profile bump.move-session-log[Codex | Claude Code]: Use when the user asks to relocate or import a session log into the current project, or to make session history and resume follow a renamed project directory in the current tool; not for merely opening or inspecting the current session log.open-session-log[Codex | Claude Code]: Open the current Claude Code or Codex session log in Emacs via agent-log. Programmatic-only helper for explicit log-opening automation; not loaded in ordinary interactive sessions.optimize-agent-instructions[Codex | Claude Code]: Optimize project agent instruction files. Use for optimize/improve/clean up/review Claude.md or making persistent CLAUDE.md or AGENTS.md guidance more concise, concrete, and reliable; not for ordinary docs or code reviews.orchestrate-review[Codex | Claude Code]: Use when coordinating two live Emacs agent sessions for a staged implementation that needs independent artifact review, fixed author/reviewer roles, unattended completion, or role reversal. Not for auditing this skill or merely explaining orchestration.org-note-conventions[Codex | Claude Code]: Use when creating or editing private or personal Org notes, regardless of directory; not package manuals, project documentation, READMEs, literate configuration, or other separately governed Org files.overnight-todos[Codex | Claude Code]: Batch-process personal org-roam TODOs when the user explicitly asks for an unattended TODO run. Classify first, perform authorized local work, and report blockers. Not for generic overnight work, auditing this skill, or creating a schedule.paste-via-kill-ring[Codex | Claude Code]: Stage text I genuinely must paste myself in the Emacs kill ring, or the macOS clipboard for a Chrome/native target. Use instead of printing text for manual copying; not when the agent can perform the authorized action directly. Credentials require the secrets workflow and must not enter the persistent kill ring by default.personalize[Codex | Claude Code]: Draft or revise text I will send or publish as myself, including email, Slack, comments, tweets, letters and reviews. Use my private voice samples and humanize’s English-prose catalogue within the drafting pass. Not for internal notes, commit messages, code, or text authored by someone else.pin-elisp-pr[Codex | Claude Code]: Pin an Emacs/Elpaca package in dotfiles to a user-requested fork, branch or PR. Preserve one effective installation owner per active profile, existing recipe settings and configuration behavior. Not for merely reviewing a PR or auditing this skill.post-push-ci[Codex | Claude Code]: Observe GitHub Actions after an authorized push, or check CI for an explicitly selected commit. Diagnose failures and, when repair is authorized, fix and verify them. Checking status alone does not authorize edits, pushes, reruns or other shared actions.pr-audit[Codex | Claude Code]: Audit a specified PR or PR-ready branch for submission or post-revision readiness. Bind the exact base/head, review commits and net changes, run safe project checks and inspect applicable CI. An audit alone does not authorize fixes, history changes or publication.profile-ai-cli-performance[Codex | Claude Code]: Benchmark or profile fresh Claude Code and Codex response latency, startup cost, and clean-versus-configured conditions. Use for configuration, hooks, MCP, skills, plugins or first-output comparisons; not to infer provider superiority or run paid requests while merely auditing this skill.proofread[Codex | Claude Code]: Proofread specified Markdown prose for spelling, grammar, punctuation and light clarity changes. Supports local aspell or explicitly scoped Gemini editing, preserving source files and reviewed suggestions; not factual review, broad rewriting or permission to upload restricted content.reconcile-contacts[Codex | Claude Code]: Compare and reconcile explicitly selected BBDB, macOS Contacts and Google contact sources, including duplicate candidates, missing records and birthday discrepancies. Use before bulk contact merges or deletions; not for one-off contact edits or a calendar-only display issue.record-decisions[Codex | Claude Code]: Extract and record architectural, algorithmic, or design decisions from the current session. Use for an explicit ADR/decision request, authorized end-of-session bookkeeping, or a significant chosen trade-off within an implementation task; not for unchosen proposals or read-only reviews.release-dotfiles[Codex | Claude Code]: Release a new version of the dotfiles repo. Use when the user says `/release-dotfiles`, asks to cut/publish/prepare a dotfiles release, bump the dotfiles version, or create the GitHub release after local profile testing. Do not use for standalone Emacs package releases; use release-package instead.release-package[Codex | Claude Code]: Release one of my Emacs packages, prepare its release notes, or audit release candidates. Use for release-package and package-release/readiness requests; preserve read-only or preparation-only scope and use release-dotfiles for the dotfiles repository.rename-package[Codex | Claude Code]: Rename an Emacs package’s files, features and API, with explicitly scoped repository, Elpaca and configuration updates. Use for an actual package rename, not ordinary prose cleanup or unrelated file/repository renames.request-review[Codex | Claude Code]: Request one independent review of a committed plan from a fresh session of the opposite CLI (Claude ↔ Codex), monitor its return, and adjudicate the findings here. Implement only when the user’s request already authorizes implementation; not a staged two-author orchestration.review-lulu-alert[Codex | Claude Code]: Review a selected LuLu outbound-firewall alert and recommend whether to allow, block or leave it unanswered. Use for an open alert or an explicitly supplied screenshot/capture; distinguish artifact review from current live state. Reviewing this skill itself uses synthetic fixtures, not live alerts.security-audit[Codex | Claude Code]: Audit development-environment secrets, supply chains, macOS security controls, and Claude/Codex agent permissions. Use for environment security reviews, secrets checks, dependency audits, or periodic security posture assessments; use specialist skills for app-binary assessments or application-code reviews.session-learning-capture[Codex | Claude Code]: Capture candidate reusable lessons from the current session when explicitly requested or invoked by its trusted configured session-end hook. Write only permitted proposal records to the central dotfiles inbox; do not promote them, implement changes, or treat quoted hooks and skill audits as capture requests.skill-audit[Codex | Claude Code]: Explicit workflow for auditing Claude/Codex skill behavior, implementation, or historical autonomous routing. Invoke only with /skill-audit in Claude or $skill-audit in Codex; never load automatically.skill-resolver[Codex | Claude Code]: Resolve Claude/Codex skill files by name when neither the user nor the current session catalog identifies the artifact. Use for unresolved name-only lookup or resolver ambiguity; not when an exact path or a catalog root-alias mapping already identifies the requested skill.symptom-check[Codex | Claude Code]: Check whether a proposed non-trivial bug fix restores the underlying invariant or only hides a symptom. Use after initial debugging and before production edits; not for trivial mechanical changes, greenfield features, or broad audits. Diagnosis and review remain read-only unless implementation is authorized.test-suite[Codex | Claude Code]: Create or expand high-value test suites for a codebase or specific module. Use when new tests, regression tests, coverage improvements, or test infrastructure are the deliverable; not for merely running an existing suite, reviewing code, or ordinary bug-fixing without a requested test deliverable.triage-personal-todos[Codex | Claude Code]: Triage a selected personal Org notes repository: assign missing 1–9 priorities when requested, preserve existing ones, assess actionability, and brief one task. Use for personal TODO prioritization or choosing what to do next; brief-only requests stay read-only. Not for Epoch tasks.try-hartree-skills[Codex | Claude Code]: Evaluate HartreeWorks third-party skills one at a time. Use when the user says “/try-hartree-skills”, asks to try Hartree skills, continue the Hartree skill queue, or record a verdict on a tested Hartree skill. Do not use for general skill discovery, local skill audits, or arbitrary third-party installs.update-log[Codex | Claude Code]: Preserve project progress through session logs and closeout bookkeeping, only when the user asks to log, close out, wrap up, or save progress, or when the deliberate end-of-session chain invokes –auto. Never invoke proactively or hand-write project logs as a substitute.verify[Codex | Claude Code]: Run an explicit user-requested criteria-driven verification loop. Use when the user asks to define success criteria, close the loop against criteria, or rigorously verify non-code outputs. Do not use for a plain live-software check owned by end-to-end, routine coding completion, or narrower audit/debug checks.walk-list[Codex | Claude Code]: Process a list file strictly item by item, sequentially or with a bounded subagent pool. Use when the user needs no skipping, batching, or peeking ahead; not for lists that can be safely read all at once.