Skip to content

Add /han-feedback skill#39

Merged
mxriverlynn merged 9 commits into
mainfrom
add-han-feedback-skill
May 29, 2026
Merged

Add /han-feedback skill#39
mxriverlynn merged 9 commits into
mainfrom
add-han-feedback-skill

Conversation

@mjansen401
Copy link
Copy Markdown
Contributor

Summary

  • Adds /han-feedback to han.core/skills/han-feedback/SKILL.md — captures structured post-session feedback on Han skills and optionally posts it as a GitHub issue to testdouble/han, behind a user-controlled sensitive-content gate
  • Adds long-form operator doc at docs/skills/han-feedback.md
  • Adds a new "Feedback" group to docs/skills/README.md (placement subject to maintainer preference — open item flagged below)
  • Adds catalog entry to root CLAUDE.md
  • Adds ## v2.8.0 entry to CHANGELOG.md

Design notes

Why han.core, not han.github: The skill's primary function is feedback capture and local file writing. GitHub posting is an optional secondary step behind an explicit user confirmation. Skills in han.github exist because GitHub interaction is their purpose (post-code-review-to-pr, update-pr-description, work-items-to-issues). This one doesn't meet that bar.

marketplace.json: The han.core description is prose-only and does not enumerate skills by name. No update needed per CONTRIBUTING.md step 6.

Skills Index group label: "Feedback" is the proposed label for the new group. If you prefer a different name or placement, happy to adjust before merge.

CHANGELOG PR link: The (PR #TBD) in the CHANGELOG will need the real PR number filled in at merge time, following the v2.7.0 pattern.

Attribution

This skill was developed by Mike Jansen and co-authored with Claude Sonnet 4.6. The SKILL.md and long-form doc reflect full plan-a-feature + plan-implementation treatment before submission — feature spec, two review agents, and a project-manager synthesis pass.

Test plan

  • Install the skill locally (han.core/skills/han-feedback/SKILL.md~/.claude/skills/han-feedback/SKILL.md or equivalent)
  • Run a session that invokes at least one han: skill, then run /han-feedback
  • Verify correct filename generated and file written to ~/.claude/han-feedback/
  • Verify duplicate detection: run /han-feedback again in the same session — it should report the existing file and stop
  • Verify sensitive-content gate: respond ambiguously to the confirmation — it should decline to post
  • Verify writing voice: no em-dashes, no prohibited vocabulary in any file

🤖 Generated with Claude Code

Captures structured post-session feedback on Han skills and routes it
to testdouble/han as GitHub issues, behind a user-controlled
sensitive-content gate. Ships one new skill (han.core), one long-form
operator doc, and updates to the skills index, CLAUDE.md catalog, and
CHANGELOG.

Co-authored-by: Claude Sonnet 4.6 <claude-sonnet-4-6@anthropic.com>
Move the han-feedback skill out of han.core into a new han.feedback
plugin (v1.0.0) that depends on han.core. The han meta-plugin does NOT
bundle han.feedback, so it is installed separately.

- Add han.feedback/.claude-plugin/plugin.json and move the skill dir
- Register han.feedback in marketplace.json; clarify the meta-plugin
  description does not bundle it
- Update docs to a five-plugin framing (README, CLAUDE.md, concepts,
  choosing-a-han-plugin, semantic-versioning, extend-han how-to,
  plugin-json-options, skills index, han-feedback long-form doc)

CHANGELOG.md intentionally left unchanged.
Add docs/how-to/provide-feedback.md covering two feedback paths:
/issue-triage for shaping ideas and vague observations into a postable
issue, and the opt-in han.feedback plugin's /han-feedback skill for
summarizing a session and posting it to testdouble/han.

Cross-reference from the how-to index, choosing-a-han-plugin, the
han-feedback and issue-triage long-form docs, and the CLAUDE.md doc map.
Group the flat 'Which path are you on?' list into five categories
(getting started, running a workflow, finding a skill or agent,
understanding how Han works, extending and contributing), and surface
the new feedback how-to under 'running a workflow'.
@mxriverlynn mxriverlynn self-assigned this May 29, 2026
@mxriverlynn mxriverlynn added the enhancement New feature or request label May 29, 2026
Make each plugin a top-level category in the skills index. han.core
keeps purpose sub-categories (planning, building, investigation,
review, discovery, conventions, operations); the smaller plugins
(han.github, han.reporting, han.feedback) are flat lists since each
covers a single purpose.
…ate, research

Make the skill boundary statements bidirectional. han-feedback already
points to code-review, investigate, and research as the skills for code
review, bug investigation, and option research; this adds the reverse
pointer to each of those three, in both SKILL.md frontmatter and the
long-form 'Do not invoke for' sections.
Add .github/ISSUE_TEMPLATE/feedback.md so anyone opening an issue gets the
same guidance the 'How to provide feedback on Han' how-to carries: be
specific, and shape the input with /issue-triage or /han-feedback before
posting. Guidance lives in HTML comments (read while composing, kept out of
the posted issue); visible headings give the author a fill-in structure.
@mxriverlynn mxriverlynn merged commit 8f47408 into main May 29, 2026
@mxriverlynn mxriverlynn deleted the add-han-feedback-skill branch May 29, 2026 21:50
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

enhancement New feature or request

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants