Eclipseic1848/CoreMind

orchestrate-child-runs

Design, implement, diagnose, or verify CoreMind Child Runs with durable identity, monotonic authority, structured concurrency, orphan audit, workspace leases, and shared projections.

contribute-coremind

Change CoreMind source within its frozen public contracts, one-way dependencies, test-first workflow, bilingual material contract, and release authorization boundary. Use when creating, changing, reviewing, or diagnosing the source and community contribution…

inspect-agent-traces

Preserve reviewable evidence through events carrying runId, eventId, sequence, and timestamp plus append-only RunState, and derive safe resume plans. Use when creating, changing, reviewing, or diagnosing the trace, runstate, and debugging capability in a…

manage-sessions

Persist multi-turn messages, fail clearly on corrupt recovery, and protect context deterministically before provider calls. Use when creating, changing, reviewing, or diagnosing the sessions and context capability in a CoreMind project.

recover-durable-runs

Design, diagnose, or verify CoreMind durable operation state, RunState atomicity, Session migration, checkpoints, effect receipts, and crash recovery without replaying uncertain side effects.

adapt-runtime-dependencies

Align CoreMind runtime dependencies behind private adapters, verify behavioral compatibility, and provide a whole-family rollback. Use when upgrading, diagnosing, or reviewing low-level model, agent-loop, tool, usage, or Session dependencies.

build-coding-agents

Build and verify a CoreMind coding agent that reproduces defects, makes minimal edits, runs target and regression tests, preserves user changes, and reports traceable evidence. Use for coding-agent configuration, implementation, evaluation, review, or…

embed-coremind-python

Drive the same Node runtime and explicit Loop over stdio JSON-RPC from Python, and register Python callables as agent tools. Use for Python SDK, protocol parity, tool bridge, or recovery work.

embed-coremind-typescript

Embed CoreMind runtime, tools, sessions, explicit Loops, evaluation, and events in Node applications through the coremind-ai facade. Use for TypeScript or JavaScript SDK integration and diagnosis.

extend-runtime-lifecycle

Design, implement, review, or diagnose CoreMind Runtime lifecycle extensions with explicit trust, minimal capability grants, timeout isolation, and truthful terminal-state preservation.

manage-providers

Inherit the provider catalog from the locked runtime dependency while keeping availability separate from real certification. Use when creating, changing, reviewing, or diagnosing the providers and models capability in a CoreMind project.

enforce-agent-permissions

Enforce ask, assisted, and full approval modes while distinguishing path-aware file tools, the Linux bash OS sandbox, and Windows shell risk boundaries. Use when creating, changing, reviewing, or diagnosing the permissions and security capability in a…

evaluate-agents

Separate runtime outcome, metrics, business evaluation, and release readiness while preventing failures from masquerading as passes. Use when creating, changing, reviewing, or diagnosing the testing, evaluation, and quality gates capability in a CoreMind…

manage-context-artifacts

Design, implement, diagnose, or verify stable context prefixes, deterministic compaction, truthful prompt-cache metrics, and safe large-output artifacts in a CoreMind project.

operate-coremind-cli

Operate or diagnose CoreMind create, run, chat, check, eval, doctor, and templates, including explicit Loop status and safe pause-resume. Use for CLI, TUI, readline, automation, exit-code, or recovery work.

build-tools

Connect deterministic business actions through built-in tools, script tools, or the stable defineTool contract. Use when creating, changing, reviewing, or diagnosing the tools and business capabilities capability in a CoreMind project.

build-verified-repair-loop

Run, diagnose, and safely adapt the offline Verified Repair Loop golden example. Use for this example's configuration, implementation, evaluation, or documented failure cases.

project-agent

Develop and verify the Verified Repair Loop CoreMind golden example. Use when adapting this example, its business rules, tools, workflow, evaluation, or failure cases.

configure-coremind

Describe agents, tools, a Workflow or explicit bounded Loop, budgets, permissions, and quality profiles in one validated coremind.yaml file. Use when creating, changing, reviewing, or diagnosing CoreMind configuration and schema.

design-workflows

Design, implement, test, or diagnose fixed workflows and explicit bounded verify-repair loops in CoreMind. Use when a task needs orchestration, convergence checks, repair limits, pause-resume, retry classification, or effect-safe recovery.

manage-checkpoints

Snapshot files before edit/write, expose diff and explicit restore, and mark unguaranteed side effects as non-reversible. Use when creating, changing, reviewing, or diagnosing the checkpoints, diffs, and restore capability in a CoreMind project.

investigate-with-bounded-research

Run, diagnose, and safely adapt the offline Bounded Research Agent golden example. Use for this example's configuration, implementation, evaluation, or documented failure cases.

project-agent

Develop and verify the Bounded Research Agent CoreMind golden example. Use when adapting this example, its business rules, tools, workflow, evaluation, or failure cases.

project-agent

Develop and verify the Contract Review Agent CoreMind golden example. Use when adapting this example, its business rules, tools, workflow, evaluation, or failure cases.

review-contracts-safely

Run, diagnose, and safely adapt the offline Contract Review Agent golden example. Use for this example's configuration, implementation, evaluation, or documented failure cases.

develop-faq-order-assistant

Run, diagnose, and safely adapt the offline FAQ and Order Assistant golden example. Use for this example's configuration, implementation, evaluation, or documented failure cases.

project-agent

Develop and verify the FAQ and Order Assistant CoreMind golden example. Use when adapting this example, its business rules, tools, workflow, evaluation, or failure cases.

analyze-sales-with-python

Run, diagnose, and safely adapt the offline Data Analysis Agent golden example. Use for this example's configuration, implementation, evaluation, or documented failure cases.

project-agent

Develop and verify the Data Analysis Agent CoreMind golden example. Use when adapting this example, its business rules, tools, workflow, evaluation, or failure cases.

design-agents

Build isolated agent instances from a focused system prompt, model options, tools, and skills. Use when creating, changing, reviewing, or diagnosing the agent construction capability in a CoreMind project.

package-agent-skills

Package reusable procedures as concise skills and inject them per agent while keeping business facts in project documentation. Use when creating, changing, reviewing, or diagnosing the skill and sop loading capability in a CoreMind project.

scaffold-coremind-projects

Generate language-aware code skeletons, tests, evaluations, bilingual documentation, SOPs, and a project skill without overwriting existing files. Use when creating, changing, reviewing, or diagnosing the templates and project guidance capability in a…