Skip to content

v5.4.0

Latest
Latest

Choose a tag to compare

View all tags
@mxriverlynn mxriverlynn released this 20 Aug 21:48
Immutable release. Only release title and notes can be modified.
han--v5.4.0
a86259a

han 5.4.0 fixes the personal-config probe that broke in every skill at once, and adds a second output style to the
communication layer. On Claude Code 2.1.228 the inline probe carried by every Project Context block stopped loading and
reported a "Contains expansion" error, which @VikiAnn reported in issue #178 and PR #184
fixed by moving the probe into a shared scripts/han-config-dir.sh that every plugin symlinks to. That one change is
why all twelve children move this release. han-communication (1.2.0) ships the han-concise output style and promotes
technical-detail separation into the canonical readability rule. han-coding (3.3.0) gates out-of-scope production
changes in /tdd and /automated-test-planning. han-plugin-builder (2.2.0) adds an authoring rule for reaching an
environment variable, and gives the guidance corpus tables of contents. han-core (3.1.1), han-documentation (1.0.1),
han-research (1.0.1), han-planning (2.2.1), han-github (2.3.1), han-reporting (2.2.1), han-feedback (2.0.2),
han-atlassian (2.3.1), and han-linear (1.1.1) carry the config-script change, the re-synced rule files, and their own
small fixes.

What's Changed

  • Han config expansion fix by @mxriverlynn in #184
  • feat(han-communication): honor the shape the reader asked for by @mxriverlynn in #186
  • feat(han-communication): add the han-concise output style and tighten technical-detail separation by @mxriverlynn in #189
  • fix(han-planning): correct two defects in the visual-material reference material by @mxriverlynn in #190
  • fix(han-coding): gate out-of-scope production changes in tdd and automated-test-planning by @mxriverlynn in #191
  • fix(han): bring skills back under Anthropic's authoring limits by @mxriverlynn in #192
  • V5.4.0 beta by @mxriverlynn in #185

Issues closed

han v5.4.0

One shared script replaces the inline config probe

Every skill carrying a Project Context block read the personal configuration directory with an inline
echo "${CLAUDE_CONFIG_DIR:-$HOME/.claude}". Claude Code 2.1.228 refused that form with a "Contains expansion" error,
so the block failed to load and the skill lost its personal configuration.
@VikiAnn reported it in issue #178, and PR #184 replaced the inline probe with
scripts/han-config-dir.sh at the repository root. Every plugin carries a symlink to that one script, and
scripts/han-config-dir.bats tests it beside it.

CLAUDE.md and CONTRIBUTING.md record the script, the symlink each plugin carries, and where its tests live.
CONTRIBUTING.md also gains an "Adding an output style" checklist, so a second style cannot ship without a long-form
doc and a plugin README scent line, and docs/templates/coverage-rule.md extends the coverage rule to output styles for
the same reason.

The readability docs describe the reworked standard

docs/readability.md covers the reworked standard: the new technical-detail-separation property, the criterion for
honoring a shape the reader asked for, and the han-concise output style beside han-readability.
docs/configuration.md, docs/concepts.md, and docs/choosing-a-han-plugin.md thread the new style and the shared
config script through the catalogs, and docs/local-development.md corrects the local setup steps.

The planning artifacts behind the readability work

docs/plans/readability-reader-format-requests/ records the specification, the implementation plan, the decision logs,
the team findings, the test plan, the scope boundary, and the discovery notes behind this release's readability changes.

A closing sweep corrected documentation drift

Commit cf60c24 fixed drift across the long-form docs: the han-atlassian dependency list in six docs, the dispatch
lists in four agent docs, the missing sibling boundaries in two more, and the plugin map in CONTRIBUTING.md.

han-communication v1.2.0

han-concise is a second output style that rolls detail up

han-communication/output-styles/han-concise.md carries the readability rule and the writing voice, plus two departures
that belong to it alone. A turn drops preamble and recap and spends no sentence carrying neither a fact nor a needed
transition. And in place of Fidelity wins, detail rolls up into the statement it supports, on the assumption that the
reader wants less than the source holds.

A floor bounds the roll-up. It keeps a fact whose loss would change what you do next, a number you will act on, and a
stated condition that bounds a claim. The self-check keeps the readability rule's eight criteria with the seventh
rewritten for the roll-up, so everything else stays in sync with han-readability. A long-form doc ships beside the
style at han-communication/docs/output-styles/han-concise.md.

Technical detail now sits after the prose, not inside it

references/readability-rule.md gains technical-detail separation as a property of the canonical standard. Prose says
what happens in plain sentences, and symbol names, file paths, flags, and snippets come after it rather than threaded
through it. han-readability and the readability-editor agent gained the same property, so all three surfaces apply
it the same way.

The shape you asked for wins

The self-check gained a criterion for matching a shape you stated: the count, the format, and the register you asked
for. When it genuinely collides with another criterion, your stated shape wins.
@mxriverlynn reported the gap in issue #177, and PR #186 fixed it. The self-check also
stopped naming a fixed number of criteria, so the standard can grow without every consuming skill going stale.

references/writing-voice.md adds stale figures of speech and foreign, Latinate, and archaic diction to the blocklist.
The three skills (edit-for-readability, explanation-guidance, readability-guidance) and references/config-rule.md
pick up the shared config-directory script.

han-coding v3.3.0

/tdd and /automated-test-planning stop at the edge of their scope

Both skills could change production code the request never asked about, on the way to making a test pass.
@mxriverlynn reported it in issue #188, and PR #191 added a scope gate to /tdd and
carried the supporting detail into tdd/references/failure-modes.md, tdd/references/test-selection.md, and
automated-test-planning/references/template.md.

/code-review gained material across five reference files: agent-dispatch.md, agent-finding-classification.md,
finding-content.md, output-verification.md, and review-checklist.md. Every skill picked up the shared
config-directory script, and the four vendored rule files (collaborative-stop-rule.md, config-rule.md,
evidence-rule.md, yagni-rule.md) re-synced with their canonical copies.

han-plugin-builder v2.2.0

Reach an environment variable through a script, not through probe text

skill-building-guidance/context-injection-commands.md documents the new rule with a results table measured on Claude
Code 2.1.234, showing which probe forms the loader accepts and which it refuses. That is the same failure Han's own
skills hit in issue #178, written down so a new skill does not repeat it. skill-reference-files.md and
troubleshooting.md both expand substantially.

The guidance corpus carries tables of contents

@kasparovabi reported in issue #182 that 111 reference files had no table of contents
and one description ran 411 characters over the 1024-character limit. PR #192 gave every reference file that lacked one
a table of contents.

han-planning v2.2.1

Two defects in the visual-material reference material are fixed. The truncated Visual Reference invariant is restored,
and the empty visual table is now bounded. @mxriverlynn reported both in issue #187,
and PR #190 fixed them.

references/operator-escalation-rule.md, references/planning-boundary-rule.md, and
references/scope-justification-rule.md gained small clarifications, the vendored rules re-synced, and all five skills
picked up the shared config-directory script.

han-core v3.1.1

/pairing resolves its config directory through the shared script instead of an inline probe, and both skills pick up
that script. The four canonical rule files (collaborative-stop-rule.md, config-rule.md, evidence-rule.md,
yagni-rule.md) gained the config-script contract and this release's readability changes, which is what the other
plugins re-synced against.

han-documentation v1.0.1

project-documentation had a generation-time sentence restored that an earlier edit dropped. Skills picked up the
shared config-directory script, and the vendored rules re-synced.

han-research v1.0.1

Every skill picked up the shared config-directory script, and the vendored rule files re-synced with their canonical
copies.

han-github v2.3.1

Every skill picked up the shared config-directory script, and the vendored rule files re-synced.
update-pr-description/references/screenshot-embed-rules.md gained material.

han-reporting v2.2.1

Every skill picked up the shared config-directory script, and the vendored rule files re-synced. The html-summary
references layout-principles.md and report-style.md both gained material.

han-feedback v2.0.2

The skill picked up the shared config-directory script, and the vendored rule files re-synced.

han-atlassian v2.3.1

Every skill picked up the shared config-directory script, and the vendored rule files re-synced. The six long-form docs
had their dependency list corrected to name all five dependencies.

han-linear v1.1.1

The skill picked up the shared config-directory script, and the vendored rule files re-synced.

Full changelog: https://github.com/testdouble/han/blob/han--v5.4.0/CHANGELOG.md#v540
Full Changelog: han--v5.3.0...han--v5.4.0

Contributors

mxriverlynn, VikiAnn, and kasparovabi
Assets 3
Loading