菜单
Stage all changes, generate a descriptive commit message from the diff, commit, and push to the current branch's upstream. Handles merge conflicts by preferring local changes for config/project files and upstream for scaffolding.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | push |
| description | Stage all changes, generate a descriptive commit message from the diff, commit, and push to the current branch's upstream. Handles merge conflicts by preferring local changes for config/project files and upstream for scaffolding. |
| user-invocable | true |
| allowed-tools | Bash, Read, Edit |
Run through the following steps in order without pausing for confirmation unless a decision point is explicitly marked as requiring one.
Delegate this to a sub-agent so the main conversation thread stays clean. Include the full text of these instructions in the agent prompt, since sub-agents cannot read skill files directly.
git --no-pager status
git --no-pager diff # unstaged changes
git --no-pager diff --cached # already-staged changes
If the working tree is completely clean and there is nothing to push, report that and stop.
Identify files that look like secrets (.env, *.pem, *_rsa, credentials.*, secrets.*). If any are present, warn the user and exclude them from staging.
If there are 10 or fewer changed files (excluding secrets), stage them individually:
git add file1 file2 ...
If there are more than 10 changed files, stage everything and then unstage secrets:
git add -A
git reset HEAD -- <secret-file-1> <secret-file-2> ...
Read the staged diff (git --no-pager diff --cached) and write a commit message that:
feat — new feature or capabilityfix — bug fixdocs — documentation onlychore — tooling, deps, config with no behavior changerefactor — restructuring without behavior changetest — test additions or fixesCo-Authored-By trailer identifying the model that generated the commit. Use the model name from your system prompt (e.g., Claude Opus 4.6, Gemini 2.5 Pro). If you can't determine the model name, use AI Assistant as a fallback.Example shape:
feat(terraform): add Cloudflare DNS module for hub provisioning
Operators can now point DNS at Cloudflare without migrating their zone.
Module is activated by dns_provider=cloudflare and requires only
CLOUDFLARE_API_TOKEN — no other provider credentials are validated.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
git --no-pager commit -m "$(cat <<'EOF'
<generated message here>
EOF
)"
Use a heredoc so multi-line messages survive the shell without escaping issues.
First, check whether the current branch has an upstream tracking branch:
git --no-pager rev-parse --abbrev-ref --symbolic-full-name @{u} 2>/dev/null
If there is no upstream, the push command later should use git push -u origin HEAD to set one.
Then pull and push:
git --no-pager pull --rebase
git push # or: git push -u origin HEAD (if no upstream)
If git pull --rebase succeeds cleanly, continue to push. If it surfaces conflicts, go to Conflict Resolution below.
When a rebase conflict appears, classify each conflicting file and resolve it non-interactively:
docs/ — spec artifacts, stories, epics, ADRs.claude/ — skills, settings, memoryCLAUDE.md, AGENTS.md — project instructions.beads/ — issue tracker stateansible/, terraform*/ — project-specific IaCsrc/ — application sourcegit checkout --ours -- <file>
git add <file>
.agents/ scaffolding files brought in from agents-core / agents-standalone upstreamscripts/specwatch.sh, scripts/specgraph.sh — shared utility scripts.claude/skills/*/references/git checkout --theirs -- <file>
git add <file>
If a file doesn't clearly fall into either bucket, stop and show the user the conflicting hunks (git --no-pager diff) and ask which side to keep before continuing.
After resolving all conflicts in a given commit, continue the rebase. Multiple commits may conflict in sequence, so loop until the rebase completes — but cap at 10 iterations to avoid runaway loops. If you hit the cap, abort the rebase (git rebase --abort) and ask the user to resolve manually.
git --no-pager rebase --continue
# If another conflict appears, resolve it the same way and continue again.
# Repeat until `rebase --continue` succeeds with no conflicts, up to 10 times.
Once the rebase finishes:
git push # or: git push -u origin HEAD (if no upstream was set)
Run git --no-pager status and git --no-pager log --oneline -3 to verify the push landed and show the user the final state. Do not prompt for confirmation — just report the result.
Pull the latest agents-core scaffolding from upstream into the current project. Use when the user wants to update their .agents/ directory, AGENTS.md, or other scaffolding files from the shared upstream repository.
Create, validate, and transition documentation artifacts (Vision, Journey, Epic, Story, Agent Spec, Spike, ADR, Persona, Runbook, Bug, Design) and their supporting docs (architecture overviews, journey maps, competitive analyses) through their lifecycle phases. Use when the user wants to write a spec, plan a feature, create an epic, add a user story, draft an ADR, start a research spike, define a persona, create a user persona, create a runbook, define a validation procedure, file a bug, report a defect, create a design, capture a wireframe, document a UI flow, sketch interaction states, update the architecture overview, document the system architecture, move an artifact to a new phase, seed an implementation plan, implement a spec, fix a bug, work on a story, or validate cross-references between artifacts. When a SPEC, STORY, or BUG comes up for implementation, always chain into the execution-tracking skill to create a tracked plan before any code is written. When execution-tracking is requested on an EPIC,
Cut a release — detect versioning context, generate a changelog from conventional commits, bump versions, and create a git tag. Use when the user says "release", "cut a release", "tag a release", "bump the version", "create a changelog", or any variation of shipping/publishing a version. This skill is intentionally generic and works across any repo — it infers context from git history and project structure rather than assuming a specific setup.
Bootstrap, install, and operate an external task-management CLI as the source of truth for agent execution tracking (instead of built-in todos). Provides the abstraction layer between spec-management intent (implementation plans and tasks) and concrete CLI commands. MUST be invoked when any implementation-tier artifact (SPEC, STORY, BUG) comes up for implementation — create a tracked plan before writing code. Optional but recommended for complex SPIKEs. For coordination-tier artifacts (EPIC, VISION, JOURNEY), spec-management must decompose into implementable children first — this skill tracks the children, not the container. Also use for standalone tasks that require backend portability, persistent progress across agent runtimes, or external supervision. Use this skill whenever the user asks to track tasks, create an implementation plan, check what to work on next, see task status, manage dependencies between work items, or close/abandon tasks — even if they don't mention "execution tracking" explicitly.
Full-lifecycle skill management — discover, install, audit, update, and detect drift for agent skills. Wraps `npx skills` when available, falls back to POSIX tooling. Use when installing shared skills, checking for upstream changes, auditing skill safety, or managing skills across projects.