@pieerry/harness-kit 3.1.2 → 3.3.1

package/.claude/plugins/staff-software-engineer/evals/test-quality.md

@@ -0,0 +1,48 @@
+# Eval: Test Quality
+
+Type: LLM-judge
+Mode: quality gate
+Threshold: weighted total >= 8.0
+
+Score each dimension 0-10. Cite test names or output lines when below 7. Weighted total = sum(score x weight%) / 100.
+
+## Rubric
+
+### Result accuracy (weight 25%)
+Does Result field (pass | fail) match exit code and counts? No misreporting.
+
+### Failure detail (weight 25%)
+If failures occurred, failing test names listed with enough context (file:line, assertion message)? If pass, Failures explicitly `none`?
+
+### Coverage of changes (weight 20%)
+Tests run cover files changed in dev phase? Gaps called out?
+
+### Command reproducibility (weight 15%)
+Command field real shell command another engineer could run as-is? Includes filters or args if used.
+
+### Duration sanity (weight 10%)
+Duration reported with unit? Plausible for suite size?
+
+### Regression risk callouts (weight 5%)
+Flakes or slow tests flagged for follow-up?
+
+## On failure (total below 8.0)
+
+Retry. Regenerate weakest section only. Max 3 attempts.
+
+## Output
+
+```json
+{
+  "scores": {
+    "result_accuracy": 0,
+    "failure_detail": 0,
+    "coverage_of_changes": 0,
+    "command_reproducibility": 0,
+    "duration_sanity": 0,
+    "regression_risk": 0
+  },
+  "weighted_total": 0.0,
+  "feedback": ["dimension: specific issue with test name or output ref"]
+}
+```

package/.claude/plugins/staff-software-engineer/guides/coding-style.md

@@ -1,14 +1,14 @@
 # Coding Style
 
-How code reads . Applies to all areas unless a `{repo}/.claude/conventions/{area}.md` says otherwise.
+How code reads. Applies to all areas unless `{repo}/.claude/conventions/{area}.md` says otherwise.
 
 ## Principles
 
-- Read before write. At least 3 similar files in the repo before producing new code.
-- Match repo conventions. Do not mix stacks (no JPA in repos using AbstractRepository, no Composition API in Vue 2 repos).
+- Read before write. At least 3 similar files in repo before producing new code.
+- Match repo conventions. No mixing stacks (no JPA in repos using AbstractRepository, no Composition API in Vue 2 repos).
 - Pragmatic over perfect. Small PRs, 1-4 files, under 100 lines ideal.
 - Test service logic always. Feature or bugfix without tests is incomplete.
-- Clean dead code when touching related files. Do not leave commented-out blocks.
+- Clean dead code when touching related files. No commented-out blocks.
 - Temp code marked: `// please remove me` so reviewers catch it.
 - Revert if issues, investigate second.
 - Defensive: null-safe, guards, edge cases.
@@ -16,7 +16,7 @@ How code reads . Applies to all areas unless a `{repo}/.claude/conventions/{area
 
 ## Never invent
 
-If a class, helper, or pattern is not in the repo, do not fabricate it.
+If class, helper, or pattern not in repo, do not fabricate.
 
 Java/Kotlin:
 ```java
@@ -28,7 +28,7 @@ Vue:
 <!-- TBD - verify with tech lead: {what is missing} -->
 ```
 
-A TODO with context beats fabricated code.
+TODO with context beats fabricated code.
 
 ## Backend defaults (Java/Kotlin/team)
 
@@ -48,4 +48,4 @@ See skills/devops/SKILL.md.
 
 ## Project overrides
 
-A repo can override any rule via `.claude/conventions/{area}.md`. See guides/conventions-override.md.
+Repo can override any rule via `.claude/conventions/{area}.md`. See guides/conventions-override.md.

package/.claude/plugins/staff-software-engineer/guides/commit-style.md

@@ -24,7 +24,7 @@ Examples:
 
 ## Body (optional)
 
-- Why, not what. Diff already shows what.
+- Why, not what. Diff shows what.
 - Reference linked tickets or related PRs.
 - Mention breaking changes explicitly with `BREAKING CHANGE:` footer.
 
@@ -32,12 +32,12 @@ Examples:
 
 - "WIP", "fixes", "more changes", "asdf"
 - Ticket missing
-- Description that just repeats the type ("fix(...): fix the bug")
+- Description repeats type ("fix(...): fix the bug")
 - Multi-line title
 
 ## Co-author
 
-When pairing or AI-assisted:
+Pairing or AI-assisted:
 
 ```
 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

package/.claude/plugins/staff-software-engineer/guides/conventions-override.md

@@ -1,10 +1,10 @@
 # Conventions Override
 
-How project-specific conventions override the plugin defaults.
+How project-specific conventions override plugin defaults.
 
 ## Path
 
-Each repo can have its own conventions file per area:
+Each repo can have own conventions file per area:
 
 ```
 {repo-root}/
@@ -16,21 +16,21 @@ Each repo can have its own conventions file per area:
  └── devops.md
 ```
 
-Only the files you need. If you do not have web work in this repo, do not create web.md.
+Only files you need. No web work in repo, no web.md.
 
 ## How it works
 
-When a skill or command generates code:
+When skill or command generates code:
 
-1. Read the plugin's default conventions from `skills/{area}/SKILL.md`.
+1. Read plugin default conventions from `skills/{area}/SKILL.md`.
 2. Check if `cwd/.claude/conventions/{area}.md` exists.
-3. If yes, read it. Rules in this file override or add to the defaults.
-4. If a rule conflicts, project wins.
-5. If no project file, use plugin defaults only.
+3. If yes, read it. Rules override or add to defaults.
+4. On conflict, project wins.
+5. No project file, use plugin defaults only.
 
-## What goes in a project conventions file
+## What goes in project conventions file
 
-The DIFFERENCES from defaults, not the full rule set. Examples:
+DIFFERENCES from defaults, not full rule set. Examples:
 
 ```markdown
 # Backend conventions for recon-service
@@ -54,15 +54,15 @@ Forbidden in this repo:
 - @Component annotation on classes (always @Service, @Repository, @Configuration)
 ```
 
-Keep it short. Document what is DIFFERENT, not what is shared with defaults.
+Keep short. Document what is DIFFERENT, not what is shared with defaults.
 
 ## Avoiding drift
 
-If a convention you wrote for the repo also applies to others, consider promoting it to the plugin defaults (skills/{area}/SKILL.md). Per-repo files should hold genuine repo-specific decisions.
+If convention you wrote for repo applies to others, consider promoting to plugin defaults (skills/{area}/SKILL.md). Per-repo files hold genuine repo-specific decisions.
 
 ## Versioning
 
-Track conventions changes in git. When a rule changes, the commit message and PR description explain why.
+Track conventions changes in git. When rule changes, commit message and PR description explain why.
 
 ## Example structure for recon-service
 

package/.claude/plugins/staff-software-engineer/guides/pipeline.md

@@ -1,13 +1,13 @@
 # Pipeline
 
-Shared rules for plan, dev, test, pr stages. Edit retry, approval, publish, and token accounting here.
+Shared rules for plan, dev, test, pr stages. Edit retry, approval, publish, token accounting here.
 
 ## Stages
 
-1. plan: generate a technical plan from the source PRP. Apply sensors, eval, retry up to 3.
-2. dev: implement the plan in code. Run lint and convention gates after each step. Retry until clean.
-3. test: run the project test suite. Capture results.
-4. pr: open a draft PR with the standard template.
+1. plan: generate technical plan from source PRP. Apply sensors, eval, retry up to 3.
+2. dev: implement plan in code. Run lint and convention gates after each step. Retry until clean.
+3. test: run project test suite. Capture results.
+4. pr: open draft PR with standard template.
 
 ## Retry policy
 
@@ -19,8 +19,8 @@ Trigger on:
 - test command exit code != 0 (dev/test only)
 
 Strategy:
-1. Read the feedback.
-2. Regenerate or fix only the failed parts.
+1. Read feedback.
+2. Regenerate or fix only failed parts.
 3. Re-run gates.
 
 Hard stop after 3 attempts: return blocker, do not proceed.
@@ -38,17 +38,21 @@ Triggers hooks/post-eval-{stage}.sh.
 
 ## Token accounting
 
-Phases tracked per feature_id (matches PRD/PRP feature_id from the PM plugin):
+Phases tracked per feature_id (matches PRD/PRP feature_id from PM plugin):
 - plan-generate, plan-validate
-- dev
-- test
-- pr
+- dev-generate, dev-validate
+- test-generate, test-validate
+- pr-generate, pr-validate
 
-Markers in outputs/.markers/{feature_id}.{phase}.{start|end}, each `{"timestamp": ISO, "session_id": ""}`. Skills write .start; hooks write .end. After each phase ends, the publish hook calls scripts/token-phase.py.
+Markers in outputs/.markers/{feature_id}.{phase}.{start|end}, each `{"timestamp": ISO, "session_id": ""}`.
+Flow per stage:
+1. Skill writes `{stage}-generate.start` before drafting artifact.
+2. Write tool fires post-write-sse.sh → writes `{stage}-generate.end` + `{stage}-validate.start`.
+3. Skill appends approval marker via Edit → fires post-eval-sse.sh → writes `{stage}-validate.end` + runs scripts/token-phase.py for both phases.
 
-Tokens land in a shared file with the PM plugin: `outputs/tokens/{feature_id}.json`. The same file collects both PM phases (prd-generate, prd-validate, prp-generate, prp-validate) and SSE phases (plan-generate, plan-validate, dev, test, pr). Totals cover the full feature lifecycle.
+Tokens land in shared file with PM plugin: `outputs/tokens/{feature_id}.json`. Same file collects PM phases (prd-generate, prd-validate, prp-generate, prp-validate) and SSE phases (plan-generate, plan-validate, dev, test, pr). Totals cover full feature lifecycle.
 
-To merge with the PM tokens file, the SSE token-phase.py writes to the same path under this plugin's outputs/tokens/, then a small step in commands/run.md syncs (or symlinks) with the PM tokens dir. For simplicity v1, SSE keeps its own outputs/tokens/{feature_id}.json. v2: merge.
+To merge with PM tokens file, SSE token-phase.py writes to same path under this plugin's outputs/tokens/, then small step in commands/run.md syncs (or symlinks) with PM tokens dir. v1: SSE keeps own outputs/tokens/{feature_id}.json. v2: merge.
 
 ## Orchestrator order
 
@@ -59,7 +63,7 @@ To merge with the PM tokens file, the SSE token-phase.py writes to the same path
 
 ## Stop conditions
 
-- 3 failed attempts on the same stage
+- 3 failed attempts on same stage
 - tests fail after dev
 - gh CLI not available for pr stage
 - missing required input after one clarification

package/.claude/plugins/staff-software-engineer/hooks/post-eval-sse.sh

@@ -0,0 +1,65 @@
+#!/bin/bash
+# Generic SSE post-eval hook. Handles plan/dev/test/pr phases:
+# - detects approved marker
+# - closes validate.end marker
+# - runs token-phase.py for both generate and validate phases
+# - appends published marker + inline tokens reference
+#
+# Registered in .claude/settings.json under PostToolUse > Edit.
+
+set -euo pipefail
+
+FILE_PATH="${CLAUDE_TOOL_FILE_PATH:-}"
+PLUGIN_DIR="$(cd "$(dirname "$0")/.." && pwd)"
+
+PHASE=""
+case "$FILE_PATH" in
+  *.claude/plugins/staff-software-engineer/outputs/plan/*.md) PHASE=plan ;;
+  *.claude/plugins/staff-software-engineer/outputs/dev/*.md)  PHASE=dev  ;;
+  *.claude/plugins/staff-software-engineer/outputs/test/*.md) PHASE=test ;;
+  *.claude/plugins/staff-software-engineer/outputs/pr/*.md)   PHASE=pr   ;;
+  *) exit 0 ;;
+esac
+
+if ! grep -q "<!-- approved:" "$FILE_PATH"; then
+  exit 0
+fi
+
+if grep -q "<!-- published:" "$FILE_PATH"; then
+  exit 0
+fi
+
+FEATURE_ID="$(basename "$FILE_PATH" .md)"
+MARKERS_DIR="$PLUGIN_DIR/outputs/.markers"
+mkdir -p "$MARKERS_DIR"
+NOW="$(date -u +%Y-%m-%dT%H:%M:%SZ)"
+
+if [ -f "$MARKERS_DIR/${FEATURE_ID}.${PHASE}-validate.start" ]; then
+  printf '{"timestamp":"%s"}\n' "$NOW" > "$MARKERS_DIR/${FEATURE_ID}.${PHASE}-validate.end"
+fi
+
+python3 "$PLUGIN_DIR/scripts/token-phase.py" \
+  --feature-id "$FEATURE_ID" \
+  --phase "${PHASE}-generate" \
+  --plugin-dir "$PLUGIN_DIR" >&2 || true
+
+python3 "$PLUGIN_DIR/scripts/token-phase.py" \
+  --feature-id "$FEATURE_ID" \
+  --phase "${PHASE}-validate" \
+  --plugin-dir "$PLUGIN_DIR" >&2 || true
+
+TOKENS_FILE="$PLUGIN_DIR/outputs/tokens/${FEATURE_ID}.json"
+if [ -f "$TOKENS_FILE" ]; then
+  TOKENS_LINE=$(python3 -c "
+import json
+with open('$TOKENS_FILE') as f:
+    d=json.load(f)
+t=d.get('totals',{})
+print(f'<!-- tokens: outputs/tokens/${FEATURE_ID}.json in={t.get(\"input\",0)} out={t.get(\"output\",0)} cache_r={t.get(\"cache_read\",0)} -->')
+")
+  printf '\n%s\n' "$TOKENS_LINE" >> "$FILE_PATH"
+fi
+
+printf '\n<!-- published: %s -->\n' "$NOW" >> "$FILE_PATH"
+echo "[hook] ${PHASE} approved + token accounting done" >&2
+exit 0

package/.claude/plugins/staff-software-engineer/hooks/post-write-sse.sh

@@ -0,0 +1,60 @@
+#!/bin/bash
+# Generic SSE post-write hook. Handles plan/dev/test/pr phases:
+# - runs sensors matching sensors/{phase}-*.md (if any)
+# - writes phase generate.end + validate.start markers
+#
+# Registered in .claude/settings.json under PostToolUse > Write.
+
+set -euo pipefail
+
+FILE_PATH="${CLAUDE_TOOL_FILE_PATH:-}"
+PLUGIN_DIR="$(cd "$(dirname "$0")/.." && pwd)"
+
+PHASE=""
+case "$FILE_PATH" in
+  *.claude/plugins/staff-software-engineer/outputs/plan/*.md) PHASE=plan ;;
+  *.claude/plugins/staff-software-engineer/outputs/dev/*.md)  PHASE=dev  ;;
+  *.claude/plugins/staff-software-engineer/outputs/test/*.md) PHASE=test ;;
+  *.claude/plugins/staff-software-engineer/outputs/pr/*.md)   PHASE=pr   ;;
+  *) exit 0 ;;
+esac
+
+if grep -q "<!-- approved:" "$FILE_PATH" 2>/dev/null; then
+  exit 0
+fi
+
+ACTIVITY="$(cd "$PLUGIN_DIR/../../.." && pwd)/.claude/scripts/activity.py"
+set_activity() { [ -x "$ACTIVITY" ] && python3 "$ACTIVITY" set "$1" "$2" 2>/dev/null || true; }
+clear_activity() { [ -x "$ACTIVITY" ] && python3 "$ACTIVITY" clear 2>/dev/null || true; }
+
+FAILURES=()
+for sensor in "$PLUGIN_DIR"/sensors/${PHASE}-*.md; do
+  [ -f "$sensor" ] || continue
+  sname="$(basename "$sensor" .md)"
+  set_activity sensor "$sname"
+  if ! python3 "$PLUGIN_DIR/scripts/sensor-runner.py" \
+        --sensor "$sensor" \
+        --artifact "$FILE_PATH" >&2; then
+    FAILURES+=("$sname")
+  fi
+  clear_activity
+done
+
+if [ ${#FAILURES[@]} -gt 0 ]; then
+  echo "[hook] ${PHASE} sensor failures: ${FAILURES[*]}" >&2
+  exit 2
+fi
+
+FEATURE_ID="$(basename "$FILE_PATH" .md)"
+MARKERS_DIR="$PLUGIN_DIR/outputs/.markers"
+mkdir -p "$MARKERS_DIR"
+NOW="$(date -u +%Y-%m-%dT%H:%M:%SZ)"
+
+if [ ! -f "$MARKERS_DIR/${FEATURE_ID}.${PHASE}-validate.start" ]; then
+  if [ -f "$MARKERS_DIR/${FEATURE_ID}.${PHASE}-generate.start" ]; then
+    printf '{"timestamp":"%s"}\n' "$NOW" > "$MARKERS_DIR/${FEATURE_ID}.${PHASE}-generate.end"
+    printf '{"timestamp":"%s","session_id":"%s"}\n' "$NOW" "${CLAUDE_SESSION_ID:-}" > "$MARKERS_DIR/${FEATURE_ID}.${PHASE}-validate.start"
+  fi
+fi
+
+exit 0

package/.claude/plugins/staff-software-engineer/sensors/code-conventions.md

@@ -3,7 +3,7 @@
 Type: deterministic (via project-specific tooling)
 Mode: hard gate
 
-Run lint and format checks for the area being touched.
+Run lint and format checks for area being touched.
 
 ## Backend (Java/Kotlin)
 
@@ -11,7 +11,7 @@ Detect via build file:
 - Maven: `mvn -q checkstyle:check` or `mvn -q spotless:check` if configured
 - Gradle: `./gradlew detekt` or `./gradlew checkstyleMain`
 
-If no linter is configured, skip and emit a warning.
+If no linter configured, skip and emit warning.
 
 ## Web (Vue/JS)
 
@@ -30,8 +30,8 @@ If no linter is configured, skip and emit a warning.
 
 ## On failure
 
-Block. Surface the linter output. Agent fixes and re-runs.
+Block. Surface linter output. Agent fixes and re-runs.
 
 ## On warning (no linter configured)
 
-Note in the output, do not block. Suggest adding the linter to the repo.
+Note in output, do not block. Suggest adding linter to repo.

package/.claude/plugins/staff-software-engineer/sensors/dev-structure.md

@@ -0,0 +1,46 @@
+# Sensor: Dev Structure
+
+Type: deterministic
+Mode: hard gate
+
+Validates dev summary doc at `outputs/dev/{feature_id}.md`.
+
+## Required sections
+
+- Summary
+- Files changed
+- Commits
+- Sensors
+- Blockers
+
+## Forbidden tokens
+
+- lorem
+- TODO without ticket reference
+- placeholder
+- XXX
+- FIXME without ticket reference
+
+## Markdown rules
+
+- exactly 1 H1 heading
+- no em-dash
+- no ASCII box-drawing
+
+## Required metadata
+
+Header must declare:
+- Source plan (path to `outputs/plan/{feature_id}.md`)
+- Branch
+- Area (backend | web | mobile | devops)
+- Date
+
+## Content rules
+
+- Files changed must list at least 1 real file path (not placeholder).
+- Commits section must list at least 1 short SHA (7+ hex chars).
+- Sensors section must report `code-conventions` and `test-coverage` outcomes (ok | failed).
+
+## On failure
+
+Block publish. Return missing sections and rule violations. Agent regenerates failed parts only.

package/.claude/plugins/staff-software-engineer/sensors/pr-structure.md

@@ -0,0 +1,49 @@
+# Sensor: PR Structure
+
+Type: deterministic
+Mode: hard gate
+
+Validates PR record at `outputs/pr/{feature_id}.md`.
+
+## Required sections
+
+- URL
+- Title
+- Draft
+- Summary
+- Test plan
+- Refs
+
+## Forbidden tokens
+
+- lorem
+- placeholder
+- TBD
+
+## Markdown rules
+
+- exactly 1 H1 heading
+- no em-dash
+- no ASCII box-drawing
+
+## Required metadata
+
+Header must declare:
+- Source plan (path to `outputs/plan/{feature_id}.md`)
+- Source dev (path to `outputs/dev/{feature_id}.md`)
+- Branch
+- Date
+
+## Content rules
+
+- URL must match `https://github.com/{owner}/{repo}/pull/{number}` shape.
+- Title must start with Conventional Commits prefix (`feat`, `fix`, `chore`, `refactor`, `docs`, `test`, `perf`, `build`, `ci`, `style`, `revert`) followed by optional scope and `:`.
+- Title length <= 70 chars.
+- Draft must be `yes` or `no`.
+- Summary must have at least 1 bullet.
+- Test plan must be markdown checklist with at least 1 item.
+- Refs section must reference both source plan and source dev paths.
+
+## On failure
+
+Block publish. Return missing sections and rule violations.

package/.claude/plugins/staff-software-engineer/sensors/test-coverage.md

@@ -3,13 +3,13 @@
 Type: heuristic
 Mode: hard gate
 
-Every feature or bugfix must include tests for the changed code.
+Every feature or bugfix must include tests for changed code.
 
 ## Check
 
-For each changed source file in the dev commit set, look for:
-- a corresponding test file (e.g., `Foo.java` -> `FooTest.java`, `bar.ts` -> `bar.test.ts`)
-- OR a test file modified in the same commit that exercises the change
+For each changed source file in dev commit set, look for:
+- corresponding test file (e.g., `Foo.java` -> `FooTest.java`, `bar.ts` -> `bar.test.ts`)
+- OR test file modified in same commit exercising the change
 
 Patterns recognized:
 - `**/test/**`, `**/tests/**`, `**/__tests__/**`, `**/spec/**`, `**/specs/**`
@@ -17,7 +17,7 @@ Patterns recognized:
 
 ## Exclusions
 
-The following do not need a test in the same commit:
+Following do not need test in same commit:
 - README and docs
 - migration files (V*.sql)
 - pure config (yaml, json)
@@ -25,4 +25,4 @@ The following do not need a test in the same commit:
 
 ## On failure
 
-Block. List the source files without tests. Agent adds the missing tests.
+Block. List source files without tests. Agent adds missing tests.

package/.claude/plugins/staff-software-engineer/sensors/test-structure.md

@@ -0,0 +1,46 @@
+# Sensor: Test Structure
+
+Type: deterministic
+Mode: hard gate
+
+Validates test report at `outputs/test/{feature_id}.md`.
+
+## Required sections
+
+- Command
+- Result
+- Counts
+- Failures
+- Duration
+
+## Forbidden tokens
+
+- lorem
+- placeholder
+- TBD
+
+## Markdown rules
+
+- exactly 1 H1 heading
+- no em-dash
+- no ASCII box-drawing
+
+## Required metadata
+
+Header must declare:
+- Source dev (path to `outputs/dev/{feature_id}.md`)
+- Framework (maven | gradle | npm | pytest | other)
+- Date
+
+## Content rules
+
+- Command must be real shell command (not placeholder).
+- Result must be `pass` or `fail`.
+- Counts must include numeric `passed` and `failed` values.
+- If Result is `fail`, Failures section must list at least 1 failing test name.
+- If Result is `pass`, Failures section must be `none`.
+- Duration must include unit (s | ms | m).
+
+## On failure
+
+Block publish. Return missing sections and rule violations.

package/.claude/plugins/staff-software-engineer/skills/backend/SKILL.md

@@ -6,11 +6,11 @@ user_invocable: false
 
 Backend conventions, team default.
 
-Project override: if `{repo}/.claude/conventions/backend.md` exists, it overrides any rule here. See guides/conventions-override.md.
+Project override: if `{repo}/.claude/conventions/backend.md` exists, overrides any rule here. See guides/conventions-override.md.
 
 ## Defaults
 
-These are illustrative defaults for a JVM-leaning backend. Override per repo via `.claude/conventions/backend.md`.
+Illustrative defaults for JVM-leaning backend. Override per repo via `.claude/conventions/backend.md`.
 
 Stack:
 - Java or Kotlin (check pom.xml or build.gradle.kts)
@@ -20,7 +20,7 @@ Stack:
 - Package prefix: match repo convention
 
 Persistence:
-- Custom JDBC AbstractRepository + SQL templates. Avoid JPA except where the repo already uses it.
+- Custom JDBC AbstractRepository + SQL templates. Avoid JPA except where repo already uses it.
 - Migrations via Flyway, naming `V{YYYYMMDD}__{description}.sql`.
 
 Mapping:
@@ -37,7 +37,7 @@ Null handling:
 
 Streams:
 - Fluent: `.stream().filter(...).findAny().orElseThrow(...)`.
-- Avoid stream-of-stream when a guard would do.
+- Avoid stream-of-stream when guard would do.
 
 Transactions:
 - `@Transactional` on service methods, not repositories.
@@ -46,12 +46,12 @@ Tests:
 - JUnit 4 + Mockito legacy; JUnit 5 + Mockito for newer.
 - `@RunWith(MockitoJUnitRunner.class)` (JUnit 4) or `@ExtendWith(MockitoExtension.class)` (JUnit 5).
 - Naming: `givenX_WhenY_ShouldZ` or `shouldDoSomethingWhenCondition`.
-- Cover positive, negative, and edge cases.
+- Cover positive, negative, edge cases.
 - `verify()` with `never()`, `times()`, arg matchers.
 
 ## Before writing code
 
-Read at least 3 similar files in the target repo to confirm:
+Read at least 3 similar files in target repo to confirm:
 - framework version
 - build tool
 - package layout
@@ -71,10 +71,10 @@ Never assume. Stack varies per repo.
 
 ## Mark gaps
 
-If you cannot find a pattern in the repo, do not invent it:
+If you cannot find pattern in repo, do not invent:
 
 ```java
 // TBD - verify with tech lead: {what you searched, what is missing}
 ```
 
-Read actual code first. A TODO with context beats fabricated code.
+Read actual code first. TODO with context beats fabricated code.

package/.claude/plugins/staff-software-engineer/skills/devops/SKILL.md

@@ -6,7 +6,7 @@ user_invocable: false
 
 DevOps conventions, team default.
 
-Project override: if `{repo}/.claude/conventions/devops.md` exists, it overrides any rule here. See guides/conventions-override.md.
+Project override: if `{repo}/.claude/conventions/devops.md` exists, overrides any rule here. See guides/conventions-override.md.
 
 ## Defaults
 
@@ -32,12 +32,12 @@ Observability:
 - Log structure: JSON, with `trace_id`, `service`, `level`, `msg`.
 
 Secrets:
-- Never commit secrets. Use a secret manager.
+- Never commit secrets. Use secret manager.
 - Pre-commit hook to scan for accidental secrets.
 
 ## Before writing code
 
-Read at least 3 similar workflows or terraform modules in the repo. Confirm:
+Read at least 3 similar workflows or terraform modules in repo. Confirm:
 - runner type (ubuntu-latest, self-hosted)
 - secret naming convention
 - environments (dev, staging, prod)