herdr-practices
An on-demand reference index for research, review, measurement, fix verification, and retrospectives; not always-on and not a mode.
- License 许可
- MIT
- 最近更新
- 2026-11-15
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- herdr-practices
- description
- An on-demand reference index for research, review, measurement, fix verification, and retrospectives; not always-on and not a mode.
02 资源文件 9
- 参考文档references/design.md5.2 KB
- 参考文档references/incident-intake.md7.4 KB
- 参考文档references/measurement.md3.1 KB
- 参考文档references/research.md3.3 KB
- 参考文档references/retrospective.md2.9 KB
- 参考文档references/review.md4.4 KB
- 参考文档references/runbooks.md8.1 KB
- 参考文档references/verify-existing-fix.md2.9 KB
- 参考文档references/writing.md4.5 KB
03 SKILL.md 正文
Herdr practices
Open only the reference needed for the current task:
- Research: explain behavior or historical rationale, with evidence and gaps.
- Design: shape costly code and architecture decisions before implementation; compare candidates only for a flagged decision.
- Runbooks: fix bugs, build features, refactor, diagnose captures, or improve measured performance.
- Review: assess a change, its decisive safety assumption, and disputed findings.
- Measurement: produce or assess a performance number.
- Verify an existing fix: compare a reported symptom before and after an existing patch.
- Incident intake: triage a bounded source batch and reproduce read-only, with access gaps and existing owners.
- Retrospective: turn demonstrated mistakes into proposals for an existing owner.
- Writing: edit briefs, reports, findings, PR bodies, owner asks, docs, and comments for clear claims and less filler.
Work directly within a bounded task. Only an authorized parent may assign a
HERDR-BRIEF through taskr and herdr agent start, using a separate tab in the
task workspace with FLEET_ROLE=worker --no-focus. The first helper in a new
task workspace uses its existing root tab. Helpers do not create their own
workspaces or worktrees, split the owner's pane, or start in-process agents.
These references grant no permission to publish, merge, operate watchers,
change services, or edit memory. Task authorization and host policy govern.
Adapted from pstack by Lauren Tan (2026); license and provenance: third_party/pstack.
04 同仓库更多 Skills
taskr
Report progress/questions/completion and orchestrate Herdr campaigns through the task ledger.
taskr
Report progress/questions/completion and orchestrate Herdr campaigns through the task ledger.
herdr-orchestration
暂无简介
herdr-fleet
暂无简介
herdr-verification
Author or maintain a project verification map using its existing driver, real user paths, and durable evidence.
herdr-orchestration
Delegate work inside Herdr (HERDR_ENV=1): the agent hierarchy, which model runs which lane, launch commands for Claude, Codex, and OpenCode workers, task layout, briefs, the taskr task ledger, waits, teardown, and recovery. Load this skill when orchestrating or sub-orchestrating, before delegating, reviewing, or starting an agent inside Herdr. The upstream `herdr` skill owns CLI syntax.