AGENT SKILL

issue-tracking

Use when agents create, select, update, comment on, close, or sync ditz issues, especially across worktrees or after uncertain or partial command results. Establish the intended tracker and use the existing CLI safely. Also use when a repo's beads/ditz instructions conflict; resolve the tracker before writing. Does not initialize trackers or migrate beads.

License 许可
AGPL-3.0
所属仓库
tedks/agent-skills
最近更新
2026-11-03
Stars
★ 0
Forks
0

01 前言 Front Matter

name
issue-tracking
description
Use when agents create, select, update, comment on, close, or sync ditz issues, especially across worktrees or after uncertain or partial command results. Establish the intended tracker and use the existing CLI safely. Also use when a repo's beads/ditz instructions conflict; resolve the tracker before writing. Does not initialize trackers or migrate beads.

02 资源文件 5

03 SKILL.md 正文

Issue tracking

Make the requested issue change and report what actually happened. Repository instructions govern the tracker; this skill is a small workflow around the existing CLI, not a new API or a replacement for project policy.

Establish the target once per context

Before mutation, identify the intended project/worktree, executable, and store:

  1. Check pwd, repository instructions, and git rev-parse --show-toplevel (or git worktree list from a bare project root). Use the assigned worktree.
  2. Resolve command -v ditz and read ditz --help=plain plus the relevant subcommand's help. Respect Nix/Bazel project tooling rules. Prefer the installed executable; if absent, report it and use an explicit project- approved install or nix run github:tedks/ditz -- <cmd> fallback. Do not silently switch binaries after a failure.
  3. Inspect relevant ~/.ditz-config settings and the project's documented storage mode. An existing ditz-metadata branch takes precedence over the configured issue directory; without that branch, the configured directory selects the filesystem store. Correlate a known issue or project identity with that store before writing. Keep private config contents out of reports.

If repo instructions, config, or the observed store disagree, resolve that specific mismatch before mutation. A missing binary, failed read, or empty listing is not permission to run init, migrate beads, create a fallback tracker, or declare that no issues exist. Do not apply ditz commands to a beads project; use its established workflow. Recheck context after changing project, executable, or config; don't repeat the full preflight per issue.

Use the current command contract

Read the relevant sections of the canonical agent recipes and JSON/retry contract. Prefer their local copies in a ditz checkout when available; compare with installed command help because that binary may predate the docs. Do not invent flags or a common JSON envelope across commands.

  • Capture IDs directly with add --ids-only, and retain the exact ID. add --id NAME creates once; an existing ID returns unchanged, even if new creation fields differ. Inspect it before relying on its contents; use set for an authorized update. This is not an upsert.
  • Feed multiline descriptions/comments through the documented stdin flags using a quoted here-document or owned file so shell metacharacters stay literal. Ditz trims outer stdin whitespace.
  • Treat ready as a snapshot and start as status/history only. Neither assigns an owner, grants a lease, nor excludes another agent. Coordinate ownership through the project's existing lane records or coordinator.
  • Capture exit status, stdout and stderr separately, including on failure. Inspect per-item successes and errors before deciding what remains. Retry only failed items whose state is known; don't replay successful closes.
  • If a mutation's receipt is lost, inspect current fields and log_events before retrying. Comments and repeated closes append events without retry deduplication. If history cannot settle the outcome, report it as unknown and resolve the uncertainty before another append.
  • Distinguish a local issue change, a metadata commit, and a successful ditz sync. Source-branch pushes do not prove metadata synchronization. Use sync within the authorized workflow and check its actual result; don't claim remote delivery from local success or an attempted push.

Report the IDs changed, any per-item failures or uncertain outcomes, and the observed sync state. Avoid dumping complete issue histories or private paths.

04 同仓库更多 Skills

tedks
tedks/agent-skills

where

Use before any find, rg, grep or Glob over a repo or a directory you do not already have an exact path for. Prints a small map of the current project (root, worktrees, depth-2 tree with sizes, the repo's own "Project structure" notes) so you search the worktree you own with a path instead of enumerating ~, ~/Projects, /, /nix or /tmp.

★ 0
tedks
tedks/agent-skills

usage

Use when a CTO or lead asks how much its lanes are spending, which session is burning quota, whether a lane tripped the token tripwire, or where the provider quota meters stand. Reads the installed token-usage counts (turns/h, cached tokens/h, average context per turn for each session; per-project/model totals; quota meters) and the tripwire's open incidents. Counts only; never message content. Does not install anything or change thresholds.

★ 0
tedks
tedks/agent-skills

stacked-prs

Detailed workflow for managing stacked PRs. Use when rebasing a PR stack, resolving merge conflicts in a stack, merging stacked PRs, or when user mentions "stack" and "PR" together.

★ 0
tedks
tedks/agent-skills

spawn-agent

Spawn AI agents in steerable tmux windows for parallel or long-running work, including transcript-preserving Codex session forks with step-scoped compaction and structured final-recap handoffs. Use to spawn interactive agents, steer running tmux agents, or fork a Codex session for a separate step.

★ 0
tedks
tedks/agent-skills

retire-agent

Preserve a private, resumable agent handoff and verify a polite exit, or transfer the handoff to a freshly spawned successor.

★ 0
tedks
tedks/agent-skills

lean-automation

Diagnose and prevent CI/deploy latency caused by automation inheriting a human's environment. Use when CI or a deploy feels slow, when auditing a repo's pipeline, when adding a dependency to a shared dev environment (flake devShell, base image, requirements file), or when writing a guard that is supposed to keep an environment lean.

★ 0