| name | flow-post-task-sync |
| description | Финализация задачи: Serena, git/GitHub, branches, worktrees. Используй для: заверши задачу, синхронизируй. EN: post-task sync, git sync, branch cleanup. |
Flow Post-Task Sync
Purpose
Leave the project in a synchronized, documented, committed state. This skill runs after Serena memory sync, not instead of it.
Workflow
- Confirm Serena memories are current. If stale, run
serena-memory-sync first.
- If
flow_post_task_state.py reports execution.agent_role=worker, do not run global sync. Return the worker JSON report to the orchestrator. Workers must not push, delete branches, install system configs, mutate project policy, or run final sync unless the orchestrator explicitly delegated that exact action.
- Resolve rldyour-flow script paths before running commands. Prefer repo-local
plugins/rldyour-flow/scripts/* when present; otherwise use ${CODEX_HOME:-$HOME/.codex}/plugins/cache/rldyour-codex/rldyour-flow/local/scripts/*. If the Stop hook provided absolute installed paths, use those.
- Read
.serena/.flow_post_task_state.json if present and run the resolved flow_post_task_state.py. Inspect branch_cleanup_state and run the resolved git_sync_audit.sh when branch/worktree cleanup is not obviously complete.
- Inspect uncommitted changes deeply. Separate source changes, docs, Serena knowledge, generated junk, runtime markers, and secrets.
- Run
instruction-docs-sync when durable project instructions may have changed and project_flow_policy.py reports instruction_docs.mode is not disabled. Keep AGENTS.md optimized for Codex and .claude/CLAUDE.md optimized for Claude Code; both are tracked normally on the main branch.
- Run applicable quality checks from project scripts and the resolved
detect_project_checks.sh.
- Commit atomically with Conventional Commits. Use separate commits for
implementation, tests/validators, docs/instructions, license/metadata,
generated artifacts, and Serena knowledge sync when that improves history
clarity or reviewability.
- Push to upstream when configured. If no upstream exists, ask before creating one.
- Agent context (
.serena/, AGENTS.md, .claude/) is tracked normally on the main branch as ordinary source. Commit it like any other source change. Runtime markers, caches, local env files, browser artifacts, and secrets remain forbidden.
- Remove merged local and remote branches/worktrees only when policy allows cleanup, the branch is not protected (
main, dev, etc.), the branch was created for this workflow, and no open PR depends on it. Advisory cleanup is reported, not forced.
- Remove
.serena/.flow_sync_marker, .serena/.flow_post_task_state.json, and .serena/.flow_blocker_ack.json only after flow_post_task_state.py reports no policy-allowed blocking reasons.
Loop Guard
Do not repeatedly delete regenerated runtime marker files. If the Stop hook repeats for the same fingerprint, the hook writes .serena/.flow_blocker_ack.json and allows Stop; report remaining blocking_reasons instead of forcing new commits.
Bootstrap-only .serena runtime files created by tool startup, such as cache entries plus runtime markers, are not meaningful project work by themselves and should not force a post-task sync loop.
Agent Context
Agent context -- .serena/memories/, .serena/project.yml, .serena/plans/, .serena/research/, .serena/newproj/, .serena/deploy/, AGENTS.md, .claude/ -- is tracked normally on the main branch as ordinary source. There is no separate agent-context branch and no agent-only overlay. Runtime markers, caches (.serena/cache/, .serena/reviews/, .serena/diagnostics/), local env files, browser artifacts, and secrets remain gitignored and forbidden from commits.
