Work

scenario Agentic repositories accumulate stale instructions quickly: README commands drift from package scripts, AGENTS.md and CLAUDE.md keep obsolete paths, examples stop matching APIs, and coding agents may trust those claims before checking the source. Evidoc turns that trust problem into a local evidence workflow.

strategy I packaged the trust check as a local workflow available from the CLI, MCP server, local web UI, and GitHub Action. Each surface compares documentation claims against repository evidence before suggesting repairs or blocking a gate.

outcome Public v0.3.2 · published on npm as @evidoc/evidoc · CLI, MCP server, Local Web UI, GitHub Action, Local Git Gate · package artifacts for CLI, core, dashboard, local app, MCP server, reports, and review log

scenario Multi-agent coding work fails when scope, privacy, review authority, and verification are implicit. A second agent is useful only if its remit is bounded and its claims are checked against the real repo. And the failures that actually cost time are rarely reasoning failures — they are runtime ones: a config value that kills the second agent at startup with no useful error, a long job that returns only an id, a flag the wrapper silently drops.

strategy I wrote the recurring practice down as skills: bound what the second agent can see or do, require review gates, and verify every claim against the repo before shipping — as a work packet where the second agent is a delegate, and as a reviewed spec and plan where it is a peer. Each runtime failure is recorded with its symptom, its actual cause, and the exact recovery, so the next session recognises it instead of rediscovering it.

outcome Public skills repository · four named workflows in two shapes — one peer mutual review, three bounded delegations · a field-tested failure log grown across 2026-07 and 2026-08 · the three delegation skills defer to the installed plugin's current contract instead of pinning a version

scenario Claude Code needed a controlled way to use a second model without leaving its local workflow or losing ownership of context. The risk was handing too much scope to another tool.

strategy I ported the official command surface to OpenCode CLI and kept job control, stop-time review, and the delegation boundary explicit. Claude Code remains responsible for files, git, verification, and final judgment.

outcome Public v0.2.0 · port of the official codex-plugin-cc surface · multi-provider second-model delegation · optional stop-time review gate

scenario Claude Code needed a bounded path to ask Grok for review or rescue work without making Grok the owner of local files or git. Manual tool switching weakens scope and evidence tracking.

strategy I ported the official command surface to Grok CLI headless mode, with job controls, a grok:grok-rescue subagent, stop-time review, and explicit boundaries. Claude Code keeps verification and shipping decisions.

outcome Public v0.2.0 · port of the official codex-plugin-cc surface · Grok CLI headless delegation · optional stop-time review gate

scenario Claude Code needed a controlled way to use a second model through Antigravity, while headless agy cannot grant reads without also allowing writes. A reviewer that can read the repository can write it too.

strategy The port keeps the command surface but makes read-only review a filesystem boundary: a disposable working-tree copy has no .git and the real tree is fingerprinted before and after. That boundary is for read-only review only — write-capable rescue runs deliberately get the real repository. Claude Code keeps files, git, and final judgment.

outcome Public v0.1.0 · port of the codex-plugin-cc surface via opencode-plugin-cc · mirror-based read-only reviews · optional stop-time review gate

scenario Codex and OpenCode are useful in different ways, but switching between them usually loses context and makes reviews harder to track. The goal was to make OpenCode a controlled collaborator inside Codex, with clear job control and a narrow privacy boundary.

strategy I exposed OpenCode through Codex MCP tools with capability checks, background job status, cancellation, and transcript boundaries. OpenCode can review or troubleshoot, while Codex keeps final control over files, git, and verification.

outcome Public v0.2.3 · eleven typed MCP tools sharing one response envelope · OpenCode review support inside Codex

scenario Grok can be useful as another engineering perspective, but switching tools manually makes scope, session history, and review ownership hard to control. The goal was to let Codex ask Grok for help while Codex still owns files, tests, git, and final judgment.

strategy I exposed Grok CLI through Codex MCP tools with capability checks, bounded run/review/rescue wrappers, session listing/export, background status/result/cancel controls, and explicit safeguards around hidden context and private runtime paths.

outcome Public Codex plugin · 0.3.0, reporting contract version 3 over a typed MCP surface with explicit workspace roots · Grok review and rescue inside Codex · bounded second-agent workflow

scenario agy ignores process cwd and is only ever told about the directory passed to --add-dir — it runs with its permission prompts skipped and is not confined to that directory — and it can report SUCCESS after a nonexistent --add-dir sends it to the wrong place. Its read and write permissions cannot be separated, and its reported outcome is independent of exit code.

strategy The plugin validates a workspace before spawn and returns a typed refusal when it cannot resolve one. Read-only reviews receive a disposable working-tree copy, and completion means resultComplete; exit code remains recorded for failure classification, not agy's verdict or completion decision.

outcome Public v0.1.0 · measured agy runtime contract with published probes · filesystem-isolated read-only reviews · restart-safe background jobs

scenario AI-coding sessions can fail because quota or account-pool status stays invisible until a provider throttles a run. A compact menu-bar signal gives users that warning before work stalls.

strategy I built a native macOS menu-bar app that polls quota and account status, then keeps the signal one click away so users can switch pools before a coding run stalls.

outcome Open-source MIT macOS menu-bar tool · built for daily AI-coding account monitoring

scenario Multi-agent developer platforms often fail at the edges: container setup, Docker instructions, fork CI, and vague diagnostics. Contributors need errors and setup paths that point to the next action.

strategy I contributed small upstream fixes that made those edges easier to operate: diagnostics surface container failures, Docker docs match current deployment, and fork CI no longer tries to publish images.

outcome 3 merged upstream PRs · diagnostics, Docker docs, and fork-safe CI

scenario Household devices such as camera batteries and Switch need separate charging rhythms and records. A device-level view keeps last-charge time and battery state together with each reminder.

strategy Device profiles pair recurring charging reminders with last-charge and battery-state records. Local notifications drive the schedule while CloudKit sync and local persistence keep the data available.

outcome Shipped to the App Store as a paid utility (spring 2025) · bilingual en/zh-Hans · has since been delisted