Creates self-contained HTML charts, tables, diagrams, and dashboards for display within a conversation. It keeps figures traceable to calculations and checks readability in light and dark themes, including narrow mobile layouts, before showing the page.
Guides code cleanup and restructuring by mapping dependencies, assessing test coverage, and planning small changes. Supports tasks such as extracting functions or renaming symbols, with checks after each step and recovery when verification fails.
Reproduce and fix LazyCodex or Codex defects in a temporary workspace. Verify the change with regression tests and real usage, then deliver an upstream Codex pull request or a LazyCodex issue containing the tested patch.
Audit a codebase with three vulnerability hunters and two engineers who build small demonstrations of suspected flaws. Verify attack paths, assess severity from demonstrated impact, and report evidence, targeted fixes, and checks to prevent recurrence.
Publishes oh-my-opencode through GitHub Actions, then checks package versions and release artifacts. Covers related packages, user-facing release summaries, and Discord announcements, reporting failures without starting code repairs during the release process.
Audit a codebase across nine areas, including architecture, tests, security, and documentation. Produces TECH_DEBT_AUDIT.md with source locations, severity ratings, estimated repair effort, and prioritized recommendations to help plan maintenance work.
Turn a broad coding request into a concrete work plan grounded in the codebase. After user approval, it documents affected files, decisions, dependencies, and verification steps so a separate worker can carry out the implementation.
Analyze all open GitHub issues and pull requests, then produce reports backed by links to specific code commits. Assess bugs, feature requests, and merge readiness without changing anything on GitHub.
Have agents with different roles challenge a proposed approach, exposing weak assumptions, missed interactions, and unnecessary complexity. Surviving findings go to a dedicated planning agent, which produces sequenced tasks with explicit success criteria.
Coordinates a named team of Codex workers with concrete responsibilities and shared progress records. It selects available collaboration tools, manages communication and separate working copies when needed, and handles integration, team archival, and cleanup after the work ends.
Finds and rewrites code by its syntax structure across 25 languages. Use it to locate matching calls or declarations, preview bulk changes, and run reusable YAML rules that check code or apply fixes.
Compare current code with the latest published npm releases to identify unpublished changes. Read actual code differences, explain their impact across shared components and agent environments, and recommend version updates for each release layer.
Checks LazyCodex and Codex CLI installations against current source code, inspecting versions, settings, plugin files, and runtime behavior. Produces an evidence-backed health report with suggested fixes while leaving the installed setup unchanged.
Find unused TypeScript code, verify references with a language server, and assign removals to parallel agents. Group changes by file, create separate commits, and check types, tests, and the build after cleanup.
Uses multiple agents with different critical roles to challenge assumptions, complexity, risks, and alternatives. Findings that withstand review go to a dedicated planner, who turns them into sequenced tasks with explicit success criteria.
Coordinate child agents on tasks that must happen in a particular order. Define dependencies, pass results through handoff files, verify the outcome, and recover individual failed tasks while reusing work that has already finished.
Search local session histories across Codex, Claude Code, OpenCode, and other coding agents. Find earlier requests, inspect transcripts and child sessions, and reconstruct past work using recorded evidence, with filters for dates, projects, platforms, and models.
Investigate crashes, hangs, incorrect responses, and other failures by testing hypotheses against observed runtime behavior. Confirm the cause, verify the fix with a reproducing test, check the real workflow, and remove temporary debugging artifacts.
Manage Git changes and investigate repository history with evidence from commands. Group related edits into focused commits, reorganize branch history when requested, trace changes, and identify the first faulty commit when a repeatable check is available.
Reviews completed changes through hands-on checks and one reviewer agent. It compares the implementation and recorded evidence with the original goal, constraints, code quality, and security requirements, then reports approval, specific blockers, or an inconclusive result.
Execute a written work plan by assigning implementation and verification to subagents. Track progress, isolate work in Git worktrees, record test evidence, and require independent verification before marking tasks complete, with optional pull request delivery.
Check the appearance and interaction states of web pages, terminal interfaces, and paginated documents. Uses captured images, light and dark themes, viewport comparisons, and independent reviews to produce evidence-backed verdicts for each checklist item.
Research recent X posts by narrowing dates, trusted accounts, and keywords. Compare posts with web search results, link to the original posts, and distinguish reported reactions from claims that lack independent confirmation.
Automate forms, clicks, screenshots, and website checks in a browser where you are already signed in. A separately launched browser also supports scraping and network capture, while login or verification steps can be handed to a person.
Is this your account? Link GitHub to bring it onto your creator page (Japanese)