Instructions file CodexOpenCode
AGENTS.md instructions for wycats/vscode-ai-plugin: This repo is both a working agent plugin and a testbed for a theory of how language shapes model behavior. Before changing skills, stances, agents, instructions, or hooks, read.
Agent
Makes controlled forward progress on a plan, reading how reality responds to each change and adjusting course without losing contact with actual workspace state.
Agent
Pre-read agent for session handoffs. Maps the relevant codebase terrain so the next working agent can orient quickly instead of discovering the landscape from scratch.
Agent
Forms falsifiable predictions about the codebase and the plan, producing hypotheses that execution will test and review will evaluate.
Agent
Bounded codebase investigator dispatched by Recon for parallel sub-investigations needing rich tools (git, terminal, exo). Returns grounded findings with sourcing, not raw data. For simple search fan-out, use Explore instead.
Agent
Adaptive codebase investigator — follows leads, uses tools interactively, and synthesizes findings. Use when each result may change what to search next. For known searches, use Explore instead.
Agent
Evaluates what happened against what was predicted, reading the gap between expectation and outcome to judge what the divergences mean and whether the result warrants confidence.
Agent
Identifies and removes slop from documents.
Skill Claude CodeCodex
Use when reviewing git diffs, local changes, or pull requests for merge-relevant risks: correctness regressions, security/privacy, data integrity, performance, reliability, tests, UX, maintainability, and deployability. Produce evidence-backed findings, inline comments when available, and a reviewer-facing summary…
Skill Claude CodeCodex
Review the conversation for threads that were started but not followed — ideas mentioned in passing, decisions deferred implicitly, tensions named but not resolved. Use when taking stock, before saving, or whenever the user wants to know what's been left on the table.
Skill Claude CodeCodex
Use when PR review or diff comments are supplied as actionable feedback to implement, including UI flows that send comments without an explicit typed request. Completing this work includes verifying the fix and resolving matching GitHub/Copilot review threads; observation-only requests should leave threads open.
Skill Claude CodeCodex
Use when drafting, updating, or reviewing a GitHub pull request title or body. Build the reviewer's intuition for the change, preserve the author's judgment and relationship with the readers, follow the repository template, and keep evidence at the strength it supports.
Skill Claude CodeCodex
Use when debugging, recon, review, or incident work has conflicting observations and needs to pause, restate what is known, check whether surfaces are comparable, and hand off to the next investigative move.
Skill Claude CodeCodex
Use when running a prepare-execute-review (PER) cycle, doing a PER workflow, or coordinating work through prepare, execute, and review subagents with explicit gates.
Skill Claude CodeCodex
Use when investigating unfamiliar code, tracing data flows, mapping architecture, or answering questions that require adaptive exploration — where the agent needs to follow leads, use tools interactively, and synthesize findings rather than just search for known terms.
Skill Claude CodeCodex
Use when drafting, revising, reviewing, or promoting RFCs and design proposals, including Exo-managed RFCs, staged RFC lifecycle work, Rust/Ember-style RFC prose, and canonization or recommendation work for related RFCs.
Skill Claude CodeCodex
End-of-day close: the user is stepping away and won't serve as a bridge. Full collaborative triage, working-style reflection, and anticipatory framing for the next day. Invoke with /session-close.
Skill Claude CodeCodex
Start a new session cycle by reconciling Project Orientation, restoring repo context, and resuming the active bet or opening vision steering when selection genuinely changed.
Skill Claude CodeCodex
End-of-cycle transition: interpret the session, triage dangling threads with the user, and prepare SESSION-TRAJECTORY.md for the next cycle. The user is still around and serves as a light bridge. Invoke with /session-rest.
Skill Claude CodeCodex
Keeps SESSION-TRAJECTORY.md warm as work progresses and offers checkpoints at natural boundaries. Saving is not stopping.
Skill Claude CodeCodex
Reorient a project by recovering its durable thesis and experienced outcome, reading live plans/tasks/PRs/substrate as candidate bets, and selecting one smallest meaningful end-to-end move with organic proof. Use when the user asks to steer or prioritize by vision, several technically valid gates compete without an…
Skill Claude CodeCodex
Use when coordinating multiple live Codex threads, maintaining a coordination heartbeat, routing work across implementation threads, inspecting cross-thread ownership or gates, or keeping a coordinated effort moving through PR, review, evidence, and human-action boundaries.
Skill Claude CodeCodex
Use when walking through code changes with the user one chunk at a time, pausing for discussion after each chunk, linking directly to code, and checking design alignment as the review progresses.
Skill Claude CodeCodex
Use when drafting or revising language that will be attributed to a user, especially public prose where editorial help must preserve the author's judgment, ownership, and relationship with the readers.