Skill3.6k estrellas del repoactualizado 6d ago

dynamic-agent-lifecycle

Historical reference for script-owned dynamic-agent lifecycle actions. Do not use from provider sessions.

Ver fuente Repositorio: claude_codex_bridge

Instalar en Claude Code

Copiar

git clone --depth 1 https://github.com/SeemSeam/claude_codex_bridge /tmp/dynamic-agent-lifecycle && cp -r /tmp/dynamic-agent-lifecycle/docs/plantree/plans/agentic-loop-workflow/drafts/agentroles.ccb_orchestrator/adapters/ccb/skills/dynamic-agent-lifecycle ~/.claude/skills/dynamic-agent-lifecycle

Después abre una sesión nueva de Claude Code; el skill carga automáticamente.

Definición

SKILL.md

# Dynamic Agent Lifecycle Reference

This reference is retained for rolepack history only. It is not projected as an
active provider skill for `agentroles.ccb_orchestrator`.

Dynamic-agent lifecycle changes are runner-owned. The orchestrator may identify
that a helper or execution role is needed, blocked, or ready for release, but
the supervisor/runner script performs all lifecycle commands and authority
writes.

## Provider Boundary

- Do not run CCB commands.
- Do not add, hide, park, resume, remove, release, or inspect agents from the
  provider session.
- Do not edit `.ccb` runtime files, provider state, panes, leases, sockets,
  mailbox records, or topology files.
- Do not choose windows, panes, or placement values.

## Reply Shape

When lifecycle information is needed, reply with semantic evidence only:

- requested role/profile and reason;
- known blockers or missing evidence;
- readiness or release recommendation;
- citations to task packet, execution contract, and supplied runtime evidence.

The runner owns command execution and treats provider replies as evidence only.

Del mismo repositorio

askSkill

Send a request to a CCB agent with `ask`.

assigned-review-chainSkill

Submit one node result to the assigned Reviewer and finish only after a bounded pass/rework chain.

bounded-work-itemSkill

Execute one scoped implementation or investigation item and return evidence without changing workflow authority.

ccb-clearSkill

Clear CCB managed agent conversation context with `ccb clear`. Use when the user writes `$ccb-clear`, `$ccb_clear`, or asks to clear/reset one or more CCB agent contexts without restarting or deleting project state.

ccb-comm-reply-recoverSkill

Diagnose and recover CCB communication and reply delivery stalls. Use when a user reports a missing CCB_REPLY, stuck ask, agent stuck busy/delivering, queued work behind an active job, cancelled/incomplete reply, empty artifact, callback not continuing, duplicate retry after success, or a CCB mailbox/communication backend that appears stuck.

ccb-compactSkill

Compact CCB managed agent conversation context with `ccb compact`. Use when the user writes `$ccb-compact`, `$ccb_compact`, or asks to summarize one or more mounted agent contexts without restarting agents or deleting project state.

ccb-configSkill

Private built-in CCB configuration skill for agentroles.ccb_self. Design, edit, validate, and prepare reloads for .ccb/ccb.config, role bindings, providers, windows, workspaces, tool windows, sidebar, and provider startup inputs. Use only inside ccb_self; non-self agents should delegate CCB config changes to ccb_self.

ccb-diagnoseSkill

Diagnose a named CCB agent by combining authoritative runtime and job lineage with deep read-only pane inspection, apply bounded recovery when evidence supports it, verify the result, and request authorization before submitting a redacted GitHub issue. Use for `$ccb_diagnose agentname`, `$ccb-diagnose agentname`, or reports that a CCB agent is stuck, disconnected, not continuing, not replying, or showing provider errors.