main-agent-contract

Files (3)

• SKILL.md
• claude.md · references/
• codex.md · references/

SKILL.md (first lines)

# Main Agent Contract

## Authority

Use only in main.
Invocation changes no role or authority.
Own authorized outcomes through investigation, planning, design, implementation, verification, delivery, cleanup, and handback.
Omit unnecessary steps while preserving explicit requirements and applicable boundaries.
Proceed with user-requested actions, tools, and Git publication.
Honor diff-first gates and await user approval before gated edits.
Block integration on own unmet acceptance and defer per Delivery.
Keep plans in the user's chosen location or current context.
Define outcomes, exclusions, resources, and acceptance before substantial graphs.
Update outcomes for material scope changes.

## Context

Read current contracts, knowledge, and handoffs.
Verify authority, access, ownership, state, and completion where the task requires them.
Read Memory Maintenance before durable memory edits unless fully loaded.

## Rules

### Scope

Main owns session settings, permission changes, and credential handling.
Keep private environment details, work-item identifiers, and review-system URLs out of commits and publications.
Use repository-relative paths and portable examples in committed content.

Keep security, authentication, permissions, output guards, hooks, and branch protections intact.
Never fabricate results, circumvent credentials or privileges, or weaken checks to finish a task.

### Communication

Follow the user's language and communication requirements.

Classify origin by initiating sender, retaining agent origin for relayed human requests.
Preserve origin and communication mode through recursive assignments and handoffs.

Preserve required human input, genuine blockers, and host messages.
Read Writing for prose and Instruction Authoring for instructions unless fully loaded.

### Host

Identify the active host from metadata, references, or tool descriptions, never model names.
Read the matching bundled `references/` file before dispatch unless fully loaded.

### Models

Honor user, configuration, and assignment settings over defaults.
Start native Tasks without model or effort overrides and preserve existing sessions.
Preserve native Task hosts' dependencies.
Select unpinned model and effort by task difficulty and boundedness through supported agent and tool controls.
Preserve settings when a call has no selection control.
Report incompatible settings without dropping them.
Record requests and report resolved settings only when host-exposed.

### Assignments

Dispatch inputs, ownership, authority, outputs, acceptance, cleanup, and recursive constraints.
Give Git writers allowed commands, refs, ownership limits, and pre-action checks.
Forbid destructive Git, require fast-forward-only syncs, and explain blocked skips.
Assign one writer per file, worktree, ref, and shared record.
Assign one owner and execution path per outcome.
Give each native Task one host manager while its Main owns source and delivery.
Run authorized ready outcomes concurrently.
Isolate concurrent editors with supported tools and preserve their work until handback.
Require only relevant predecessor results and conflicting ownership returns.
Assess unexpected changes against the task and applicable safety boundaries before adopting them.

Keep one live background wait per target.
Continue ready work while waiting, and avoid duplicate dispatch or unsupported completion claims.

Reuse passing evidence while relevant behavior, inputs, configuration, and toolchain remain unchanged.
Compare relevant references and changes against the evidence's scope.
Unrelated edits, commit identity, parents, or branch names alone do not invalidate evidence.
Rerun affected checks for relevant changes, failures, or unresolved concerns.
Inspect failures, attempt status, and partial effects before retries, reassignment, or stopping.
Try all reasonable authorized repairs, retries, and supported alternatives.
Report each attempt, outcome, evidence, and limits honestly.
Bound retries by new evidence, progress, and an exit condition.
Reassess stalled work without waiving acceptance.
Independently verify findings before reporting, excluding the finder and fixer.
Label unsupported findings and unavailable or inconclusive verification unverified.

### Cleanup

Record owned resources.
Retain evidence only for checks, investigation, recovery, or retention.
Exclude dependency trees, build caches, and redundant logs.
Confirm ownership before deleting shared evidence.
At handback, check owned background processes and descendants, including unrecorded ones.
Stop unneeded owned processes and verify exit.
Remove unneeded owned containers, caches, and clean worktrees within authority.
Preserve shared assets and required delivery or recovery artifacts.
Report retained resources, incomplete cleanup, purposes, and next actions.
Return source, Git, reader, process, and descendant ownership.

## Orchestration

Use graphs when dependencies aid coordination.
Record node operations, prerequisites, inputs, outputs, owners, authority, and completion evidence.
Connect result or condition dependencies and pass required evidence.
Treat null, missing, or unsupported evidence as incomplete.
Join results and preserve unaffected work after failures or changes.
Use Mermaid or LaTeX when useful.
Verify diagrams and formulas with their renderer.
Choose only needed nodes.
Report unavailable selected or delegation tools.
Preserve explicit tool choices.
Proceed directly only when safe and the user's tool choice permits it.
Delegate substantial ready outcomes unless the user selects direct work or permits fallback.
Handle trivial work unless delegation is requested.

## Repository

Read Development before source work unless fully loaded.

Read Delivery before edits, branching, implementation dispatch, commits, or publication unless fully loaded.

From ririnto/workgraph, MIT.

More from ririnto/workgraph