Standup and changelog¶

Standup summary¶

st standup                 # last 24 hours (default)
st standup --hours 48      # look back further
st standup --all           # include all stacks
st standup --json          # raw activity as JSON
st standup --ci            # also check CI for selected branches
st standup --ci --json     # CI results plus machine-readable signal states

Human-readable standup output shows progress while stax collects Git/forge/Jira context. Use --json for machine-readable output without progress lines.

Standup summary

Shows merged PRs, opened PRs, recent pushes, authored reviews, and items needing attention. Works with GitHub, GitLab, and Gitea.

GitHub authored reviews use one time-bounded GraphQL request capped at 100 results, followed by local repository and time filtering. GitLab and Gitea report this signal as unsupported; stax does not scan every merge or pull request. API and authentication failures are reported as unavailable instead of being presented as a successful empty result.

CI is deliberately off by default, so ordinary standup collection adds no per-branch CI latency. Pass --ci to check only the selected current-stack branches with bounded concurrency. --all --ci intentionally widens that scope to all tracked branches.

JSON keeps the compatibility arrays reviews_given and needs_attention.ci_failing. Read the companion signals.reviews_given and signals.ci_failing entries before interpreting an empty array: statuses are available, unsupported, unavailable, or not_requested, with a reason when useful.

AI standup summary¶

Generate a concise spoken-style summary using your configured AI agent:

st standup --ai
st standup --ai --hours 48
st standup --ai --agent claude
st standup --ai --style slack  # Slack-ready Yesterday/Today bullets
st standup --ai --jit       # add Jira context via jit

Uses the agent configured under [ai] in ~/.config/stax/config.toml (same agent as st gen / st generate --pr-body). Override per-run with --agent.

With --jit, standup inspects your current Jira sprint via the jit CLI and feeds the AI two extra signals:

tickets with PRs already in flight
likely next backlog tickets without PRs

The summary is word-wrapped into a card fit to your terminal width:

  ✓ Generating standup summary with codex        4.1s

  ╭──────────────────────────────────────────────────────────────────╮
  │                                                                  │
  │  Yesterday I finished the billing webhook retry fix and split    │
  │  the reporting dashboard cleanup into two PRs so review stays    │
  │  small. I also opened a third PR to speed up integration tests   │
  │  by caching seed data, and all three are now in review. Today    │
  │  I'm handling review feedback and preparing the next analytics   │
  │  task.                                                           │
  │                                                                  │
  ╰──────────────────────────────────────────────────────────────────╯

Key phrases are highlighted: completed work in green, new work in cyan, reviews in blue, upcoming tasks in yellow.

For a copy-ready Slack update, use the Slack style:

st standup --ai --style slack

It prints plain text with the same shape as a team standup thread, carrying unfinished branch, PR, or Jira work into Today when the activity shows something is still in flight:

Yesterday:
• finished the billing webhook retry fix
• opened the reporting dashboard cleanup for review

Today:
• handle review feedback
• prepare the next analytics task

Output formats¶

st standup --ai               # spinner + colored card (default)
st standup --ai --style slack # Slack-ready Yesterday/Today bullets
st standup --ai --plain-text  # raw text, pipe-friendly
st standup --ai --json        # {"summary": "..."}

Progress feedback is shown for the default card and Slack styles while context collection and AI generation run. Use --ai --plain-text or --ai --json when stdout must contain only the generated summary.

Prerequisites¶

An AI agent installed on PATH: claude, codex, gemini, opencode, or pi
For --jit: jit on PATH
Agent configured in ~/.config/stax/config.toml:

[ai]
agent = "claude"   # or "codex", "gemini", "opencode", "pi"

Or pass --agent directly.

Changelog generation¶

st changelog                   # auto-detect last tag → HEAD
st changelog v1.0.0            # explicit from ref → HEAD
st changelog v1.0.0 v2.0.0     # between two refs
st changelog abc123 def456     # between two commits
st changelog find              # fuzzy-find commits interactively
st changelog find "auth fix"   # fuzzy-find commits in the default range
st changelog --find "auth fix" # flag form for scripts

PR numbers are extracted from squash-merge commit messages like (#123).

find / --find searches commit messages in the same range used by st changelog. Use positional refs with the flag form, for example st changelog v1.0.0 HEAD --find "auth fix". Add --json with a query for scriptable output.

Monorepo tag prefix¶

With platform-scoped tags like release/ios/v1.2.0, pick the latest tag matching a prefix:

st changelog --tag-prefix release/ios
st changelog --tag-prefix release/android --json

Path filtering¶

st changelog v1.0.0 --path apps/frontend
st changelog v1.0.0 --path packages/shared-utils
st changelog find "auth fix" --path apps/frontend

JSON output¶

st changelog v1.0.0 --json
st changelog --json              # auto-resolved tag appears in "resolved_from"