appydave-tools 0.76.3 → 0.76.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: d2b65bd11cb0b4e0a0c9b1718a9ebb262ad6c998167bfdc8f533ddca754ee850
-  data.tar.gz: e45a7a56eb7f77249911ea874037c002829cfcca1f8f5cfd8553b60d12e88898
+  metadata.gz: 63d45ba7b83fca823b63a907473c424701e50c117ccac57cccdc1d31f2ae277c
+  data.tar.gz: 1dff3941a4d00ecc51b84b5b955c7640c955ee0037f52040e4469f5d9b007c4f
 SHA512:
-  metadata.gz: feb57bd1aca715ae2b6b3a587e762e024457b4d14f1172069d1afd19e953c6c432567d9c76c819543f740acb757c53bb36fa7bc839e4e7c8706f13435e94bfb0
-  data.tar.gz: eba77632d7dad881243726e122d8da5546040910b2851c36a9b751d7310da9ca3d1a4279571baa31b908bb1ff398436253bed4f7e52907adeeb6e4cfe1b1f69a
+  metadata.gz: fa0a3525cae67422cde5aa4cc880362fbeb24b6ee5d914ece224965daa07ce61878da662aec431edf3c10d9feae53cc8ec3e383bc10fa6c8267f92ade321a8ff
+  data.tar.gz: d36d104f67799d434b592effa0e6a029b92ec1e1005216c68269ea0637f67c5fdd2867505c42b900f497d16218a425f480503c6d18cffc6afc3d64b3c9313774

data/CHANGELOG.md

@@ -1,3 +1,10 @@
+## [0.76.3](https://github.com/appydave/appydave-tools/compare/v0.76.2...v0.76.3) (2026-03-19)
+
+
+### Bug Fixes
+
+* add field isolation assertions to update_spec non-updated fields ([da678f6](https://github.com/appydave/appydave-tools/commit/da678f68f5ff0384020639069b459d98e0558968))
+
 ## [0.76.2](https://github.com/appydave/appydave-tools/compare/v0.76.1...v0.76.2) (2026-03-19)
 
 

data/README.md

@@ -43,7 +43,6 @@ ad_config -c
 This creates empty configuration files at `~/.config/appydave/`:
 - `settings.json` - Paths and preferences
 - `channels.json` - YouTube channel definitions
-- `youtube-automation.json` - Automation workflows
 
 **2. Option A: Copy example files**
 
@@ -276,38 +275,6 @@ prompt_tools completion -f prompt_template.md -k topic=Ruby,style=tutorial -c
 
 ---
 
-### ⚡ YouTube Automation *(Internal/Experimental)*
-
-**The problem:** Video content creation workflows involve multiple steps: research → scripting → production.
-
-**The solution:** Run predefined prompt sequences against OpenAI API to automate research and content generation steps.
-
-```bash
-# Run automation sequence (requires configuration)
-youtube_automation -s 01-1
-
-# With debug output
-youtube_automation -s 01-1 -d
-```
-
-**What it does:**
-- Loads sequence configuration from `~/.config/appydave/youtube_automation.json`
-- Reads prompt templates from Dropbox (`_common/raw_prompts/`)
-- Executes OpenAI API calls for each sequence step
-- Saves responses to output files
-
-**Configuration required:**
-- Sequence definitions in `youtube_automation.json`
-- Prompt template files in configured Dropbox path
-- `OPENAI_ACCESS_TOKEN` environment variable
-
-**Current status:** ⚠️ **Internal tool** - Hardcoded Dropbox paths, uses deprecated Completion API, not documented for external use.
-
-**Relationship to other tools:** This is separate from **Move Images** tool (which organizes downloaded images into video project asset folders).
-
-**Use cases:** Automated content research, script outline generation, multi-step prompt workflows.
-
----
 
 ### ⚙️ Configuration Manager
 
@@ -338,7 +305,6 @@ ad_config -p
 |------|---------|-------------------------|
 | `settings.json` | Paths and preferences (video-projects-root, download folders, etc.) | ✅ Yes (after removing personal paths) |
 | `channels.json` | YouTube channel definitions (code, name, youtube_handle, locations) | ✅ Yes (share structure, customize paths locally) |
-| `youtube_automation.json` | Automation workflow configurations | ✅ Yes (if no sensitive data) |
 | `.env` | **Secrets and API keys** | ❌ **NEVER** (gitignored) |
 
 **Key Settings:**

data/docs/planning/BACKLOG.md

@@ -1,30 +1,20 @@
 # Project Backlog — AppyDave Tools
 
-**Last updated**: 2026-03-19 (bugfix-and-security assessment + 4 new items from quality audit)
-**Total**: 27 | Pending: 17 | Done: 10 | Deferred: 0 | Rejected: 0
+**Last updated**: 2026-03-19 (final-test-gaps campaign complete; B031/B032/B033 added from quality audit)
+**Total**: 33 | Pending: 11 | Done: 22 | Deferred: 0 | Rejected: 0
 
 ---
 
 ## Pending
 
-### High Priority
-- [ ] B024 — Tests: add configure_ssl_options unit tests to s3_operations_spec + share_operations_spec (protects B017 security fix) | Priority: high
-- [ ] B015 — BUG-2: FileCollector uses FileUtils.cd without ensure (process dir not restored on exception) | Priority: high
-- [x] B006 — BUG-1: Jump CLI get/remove key lookup | Completed: verified fixed 2026-03-19, regression spec added
-
 ### Medium Priority
-- [ ] B022 — Tests: expand cli_spec.rb with functional tests (-i, -e, -f, -o flags, exit codes) | Priority: medium
-- [ ] B026 — Tests: add determine_range edge cases (b00, b9, a40) to sync_from_ssd_spec + manifest_generator_spec | Priority: medium
-- [ ] B027 — Tests: strengthen gpt_context no-args spec to verify file collection actually stops | Priority: medium
-- [ ] B023 — Tests: add file_collector_spec coverage for JSON format, aider format, error paths | Priority: medium
-- [ ] B018 — Tests: add specs for Jump Commands::Remove, Commands::Add, Commands::Update | Priority: medium
-- [ ] B019 — Fix: remove debug puts @working_directory from gpt_context/file_collector.rb | Priority: medium
 - [ ] B001 — FR-1: GPT Context token counting | Priority: medium
 - [ ] B012 — Arch: add integration tests for brand resolution end-to-end | Priority: medium
 
-- [ ] B025 — Fix: stale comment in sync_from_ssd.rb line 173 (says 60-69, should say b50-b99) | Priority: low
-
 ### Low Priority
+- [ ] B031 — Tests: add_spec.rb assert `type` field in location data integrity test | Priority: low
+- [ ] B032 — Tests: cli_spec.rb add subprocess test for `-f json` flag | Priority: low
+- [ ] B033 — Fix: file_collector.rb line 19 return `''` directly when working_directory doesn't exist | Priority: low
 - [ ] B007 — Performance: parallel git/S3 status checks for dam list | Priority: low
 - [ ] B008 — Performance: cache git/S3 status with 5-min TTL | Priority: low
 - [ ] B009 — UX: progress indicators for dam operations > 5s | Priority: low
@@ -45,6 +35,19 @@
 - [x] B016 — BUG-3: ManifestGenerator + SyncFromSsd incompatible SSD range strings | Completed: bugfix-and-security (2026-03-19)
 - [x] B017 — Security: ssl_verify_peer disabled unconditionally in S3Operations + ShareOperations + S3Scanner | Completed: bugfix-and-security (2026-03-19)
 - [x] B021 — Fix: gpt_context no-args guard had dead format.nil? condition | Completed: bugfix-and-security (2026-03-19)
+- [x] B024 — Tests: configure_ssl_options unit tests (protects B017 fix) | Completed: test-coverage-gaps (2026-03-19)
+- [x] B022 — Tests: functional cli_spec tests for -i -e -f -o flags | Completed: test-coverage-gaps (2026-03-19)
+- [x] B026 — Tests: determine_range edge cases (b00, b9, a40) | Completed: test-coverage-gaps (2026-03-19)
+- [x] B027 — Tests: gpt_context no-args spec verifies collection stops | Completed: test-coverage-gaps (2026-03-19)
+- [x] B025 — Fix: stale comment sync_from_ssd.rb:173 | Completed: test-coverage-gaps (2026-03-19)
+- [x] B018 — Tests: Jump Commands::Remove, Add, Update specs | Completed: test-coverage-gaps (2026-03-19)
+- [x] B015 — BUG-2: FileCollector FileUtils.cd without ensure | Completed: already fixed in commit 13d5f87; closed without campaign 2026-03-19
+- [x] B019 — Fix: remove debug puts @working_directory from file_collector.rb | Completed: already removed prior to test-coverage-gaps campaign; closed without campaign 2026-03-19
+- [x] B006 — BUG-1: Jump CLI get/remove key lookup | Completed: verified fixed 2026-03-19, regression spec added
+- [x] B023 — Tests: file_collector_spec json/aider/error path coverage | Completed: final-test-gaps (2026-03-19)
+- [x] B028 — Tests: cli_spec file body content assertions in -i/-e tests | Completed: final-test-gaps (2026-03-19)
+- [x] B029 — Tests: add_spec validate all returned location data fields | Completed: final-test-gaps (2026-03-19)
+- [x] B030 — Tests: update_spec verify non-updated fields unchanged | Completed: final-test-gaps (2026-03-19)
 
 ---
 
@@ -64,40 +67,6 @@
 
 ---
 
-### B002 — FR-2: GPT Context AI-Friendly Help System
-
-**Spec:** `docs/specs/fr-002-gpt-context-help-system.md`
-
-**User Story**: As an AI agent using GPT Context via skills, I want structured, comprehensive help output so I can understand all options and use the tool correctly.
-
-**Notes:** Full spec exists. High priority — enables AI skill integration.
-
----
-
-### B006 — BUG-1: Jump CLI get/remove Key Lookup Bug
-
-**Priority:** High — but verify live first
-
-**⚠️ Three-lens audit (2026-03-19) found:** Static analysis of `Jump::Config#find`, `Config#remove`, and `Search#get` shows correct dual-key guards. Bug may be environmental (stale config file format) or already fixed in a prior commit. **Run `bin/jump.rb get <key>` live before writing any code.**
-
-**Steps to Reproduce:**
-```bash
-bin/jump.rb search awb-team   # ✅ finds entry
-bin/jump.rb get awb-team      # ❌ "Location not found"  ← confirm this still fails
-bin/jump.rb remove test-minimal --force  # ❌ "No locations found."
-```
-
-**If bug confirmed:** Failure site is in `Jump::Config#find` (`loc.key == key`) or the memoization of `@locations` — NOT in `name_manager/`. The `Commands::Remove#run` → `config.find(key)` → `locations.find { |loc| loc.key == key }` path is the call chain to trace.
-
-**Acceptance Criteria:**
-- [ ] Confirm bug still exists live (if not, write regression spec and close)
-- [ ] `get <key>` finds entry when `search <key>` finds it
-- [ ] `remove <key>` finds entry when `search <key>` finds it
-- [ ] Regression tests for key lookup consistency
-- [ ] Root cause documented in commit message
-
----
-
 ### B007 — Performance: Parallel git/S3 Status Checks
 
 **Context:** `dam list appydave` with 13 projects takes ~26s (sequential git+S3 check per project).
@@ -146,47 +115,6 @@ bin/jump.rb remove test-minimal --force  # ❌ "No locations found."
 
 ---
 
-### B015 — BUG-2: FileCollector FileUtils.cd Without ensure
-
-**Context:** `lib/appydave/tools/gpt_context/file_collector.rb` line 20 calls `FileUtils.cd(@working_directory)`. If an exception fires inside `build`, the process working directory is never restored to `Dir.home`. Affects any subsequent operation in the same process.
-**Fix:** Wrap in block form: `FileUtils.cd(@working_directory) { ... }` — Ruby handles restore automatically. Or add `ensure FileUtils.cd(Dir.home)`.
-**Blocker for:** FR-2 — fix this before adding more code paths to `file_collector.rb`.
-**Source:** Code quality audit 2026-03-19, MAJOR issue #2.
-
----
-
-### B016 — BUG-3: ManifestGenerator vs SyncFromSsd Incompatible Range Strings
-
-**Context:** `ManifestGenerator.determine_range('b65')` returns `"b50-b99"`. `SyncFromSsd.determine_range('b65')` returns `"60-69"`. Both are used to construct SSD archive folder paths. Projects archived via SSD sync won't be found by the manifest generator, and vice versa. A `find_ssd_project_path` fallback scan may paper over this in practice.
-**Fix:** Standardise both methods on one format. Decide which format matches actual SSD folder structure on disk, then update the other method to match.
-**Source:** Code quality audit 2026-03-19, MAJOR issue #1.
-
----
-
-### B017 — Security: ssl_verify_peer Disabled in S3Operations + ShareOperations
-
-**Context:** `lib/appydave/tools/dam/s3_operations.rb` lines 110-113 and `share_operations.rb` lines 97-100 set `ssl_verify_peer: false` unconditionally. Comment claims "safe for AWS S3" — this is incorrect. Disabling peer verification removes MITM protection on all S3 operations including AWS credential transmission.
-**Fix:** Remove the `ssl_verify_peer: false` override entirely. AWS SDK handles SSL correctly by default. If there was a historical reason (corporate proxy, dev cert issue), document it and scope to `ENV['AWS_SKIP_SSL'] == 'true'` only.
-**Source:** Code quality audit 2026-03-19, MAJOR issue #3.
-
----
-
-### B018 — Tests: Jump Commands Layer Has No Dedicated Specs
-
-**Context:** `Commands::Remove`, `Commands::Add`, `Commands::Update` have zero unit specs. The CLI spec only tests auto-regenerate side effects. `--force` guard, error codes, suggestion-on-not-found logic, and key nil-handling in these commands are entirely untested.
-**Fix:** Add `spec/appydave/tools/jump/commands/remove_spec.rb`, `add_spec.rb`, `update_spec.rb`. Use `JumpTestLocations` factory + `with jump filesystem` context.
-**Source:** Test quality audit 2026-03-19, RISK-1.
-
----
-
-### B019 — Fix: Remove Debug puts From file_collector.rb
-
-**Context:** `lib/appydave/tools/gpt_context/file_collector.rb` line 15: `puts @working_directory`. Prints working directory on every `gpt_context` invocation. Pollutes captured output. Must be removed before FR-2 adds help system output.
-**Fix:** Delete line 15.
-**Source:** Code quality audit 2026-03-19, MINOR issue #4.
-
----
-
 ### B020 — Arch: Split S3Operations (1,030 lines)
 
 **Context:** `S3Operations` handles upload, download, status, cleanup, archive to SSD, MD5 comparison, content type detection, file formatting, and time formatting — all in one class with direct `puts` throughout. No result objects returned from upload/download/cleanup (void methods with side effects).

data/docs/planning/final-test-gaps/IMPLEMENTATION_PLAN.md

@@ -5,15 +5,17 @@
 **Target**: All 4 work units complete; 817+ examples passing; rubocop clean; no regressions; regression catch rate meaningfully above 55%
 
 ## Summary
-- Total: 4 | Complete: 0 | In Progress: 4 | Pending: 0 | Failed: 0
+- Total: 4 | Complete: 4 | In Progress: 0 | Pending: 0 | Failed: 0
 
 ## Pending
 
 ## In Progress
-- [~] fix-b023 — file_collector_spec: add json format, aider format, and error path tests
-- [~] fix-b028 — cli_spec: add file body content assertions to -i and -e tests
-- [~] fix-b029 — add_spec: validate ALL returned location data fields match input attrs (path, jump, tags, description)
-- [~] fix-b030 — update_spec: verify non-updated fields are unchanged on both updated record and sibling records
+
+## Complete
+- [x] fix-b023 — file_collector_spec: +10 examples (5 json, 4 aider, 1 error path). Key fix: json exclusion test needed `exclude_patterns: ['excluded/**/*']` not `[]`. Nonexistent dir test used non-matching glob to avoid false positives. 830 examples, 85.92% coverage.
+- [x] fix-b028 — cli_spec: +6 body assertions across -i (3) and -e (3) blocks. No new example count (assertions added to existing `it` blocks). 830 examples, v0.76.3 released.
+- [x] fix-b029 — add_spec: +1 example asserting path/jump/tags/description. Confirmed `location.to_h` uses symbol keys and `.compact`. 830 examples.
+- [x] fix-b030 — update_spec: +2 examples (non-updated fields on updated record; sibling record field isolation). 11→13 examples. 830 examples.
 
 ## Complete
 

data/docs/planning/final-test-gaps/assessment.md

@@ -0,0 +1,74 @@
+# Assessment: final-test-gaps
+
+**Campaign**: final-test-gaps
+**Date**: 2026-03-19 → 2026-03-19
+**Results**: 4 complete, 0 failed
+**Version shipped**: v0.76.3
+**Quality audit**: code-quality-audit + test-quality-audit run post-campaign
+
+---
+
+## Results Summary
+
+| Work Unit | Examples Added | Notes |
+|-----------|---------------|-------|
+| fix-b023 | +10 | file_collector_spec: 5 json + 4 aider + 1 error path. Agent fixed json exclusion test (needed `exclude_patterns: ['excluded/**/*']` not `[]`) |
+| fix-b028 | +0 new examples, +6 assertions | Body content assertions added to existing -i/-e `it` blocks in cli_spec |
+| fix-b029 | +1 | add_spec: location data integrity (path/jump/tags/description) |
+| fix-b030 | +2 | update_spec: non-updated fields on updated record + sibling record isolation |
+
+**817 → 830 examples (+13). 85.61% → 85.92% coverage.**
+
+---
+
+## What Worked Well
+
+- **B023 agent self-corrected the json exclusion test.** Plan specified `exclude_patterns: []` for the json exclusion test — agent recognised this would not actually test exclusion and changed to `exclude_patterns: ['excluded/**/*']`. Agents reading source before writing improved test quality.
+- **B023 agent handled the nonexistent-dir edge case correctly.** Used a pattern that matches nothing in CWD (`['**/*.nonexistent_xyz_12345']`) to avoid false positives when `build_formats` falls through to current directory.
+- **Parallel wave, zero merge conflicts.** 4 agents, all different files.
+- **Regression catch rate measurably improved.** C (55%) → B (70-75%). All major gaps from the prior audit are closed.
+
+---
+
+## What Didn't Work
+
+**Code quality MINOR — `file_collector.rb:19` silent collection from CWD:**
+When `working_directory` doesn't exist, `build_formats` runs without `FileUtils.cd`, meaning `Dir.glob` runs in the current process directory. The spec confirms "returns empty string" only because the test uses a non-matching glob — not because the code guarantees it.
+
+**Test gap — `type` field missing from add_spec data integrity test:**
+`valid_attrs` includes `type: 'tool'` but the new B029 integrity test doesn't assert `location[:type]`. Field mapping bug for `type` would still pass silently.
+
+**Test gap — no CLI-level test for `-f json` or `-f aider` format flags:**
+These formats are unit-tested in file_collector_spec but not via subprocess in cli_spec. A regression in CLI flag parsing for these formats would pass.
+
+---
+
+## Key Learnings — Application
+
+- **`build_formats` fallthrough is a silent failure mode.** When `Dir.exist?` is false, the code returns `build_formats` result — not `''`. The guard at line 19 should probably return `''` directly rather than calling `build_formats` without cd.
+- **`exclude_patterns: []` tests exclusion vacuously.** An empty exclude list means no exclusion test is happening. Always use a real pattern when testing that something is excluded.
+- **`build_aider` embeds unsanitised prompt and file paths.** Prompts with quotes or paths with spaces produce malformed aider command output. Low severity now, worth capturing.
+- **Agent pre-read + source verification pattern is working.** Both b029 and b030 agents read the source files and confirmed field accessor names before writing assertions — zero failures from wrong field names.
+
+---
+
+## New Backlog Items from Quality Audit
+
+- **B031** — Tests: add_spec.rb assert `type` field in data integrity test | Priority: low
+- **B032** — Tests: cli_spec.rb add subprocess test for `-f json` flag | Priority: low
+- **B033** — Fix: file_collector.rb line 19 — return `''` directly when working_directory doesn't exist (don't delegate to build_formats) | Priority: low
+
+---
+
+## Suggestions for Next Campaign
+
+**Test debt is largely cleared.** The suite is at B grade (70-75% regression catch rate). The remaining gaps (B031/B032/B033) are low severity.
+
+**Recommended next move: architectural work.**
+
+- **B011** — Extract VatCLI business logic from `bin/dam` (1,600 lines). This is the prerequisite for any parallelism or performance work. 20+ rubocop-disable comments are the symptom.
+- **B020** — Split `S3Operations` (1,030 lines). Required before B007 (parallel S3 checks) can be built cleanly.
+
+These are larger than any prior campaign. Recommend scoping carefully — one of the two per campaign, not both. B011 first (it's the CLI entry point and larger structural problem).
+
+Or, if David wants a quick win: B031/B032/B033 as a micro-campaign (3 test items, 1 production fix, one wave).

data/docs/planning/micro-cleanup/AGENTS.md

@@ -0,0 +1,181 @@
+# AGENTS.md — micro-cleanup
+
+> Inherited from final-test-gaps AGENTS.md. Self-contained.
+> Last updated: 2026-03-19
+
+---
+
+## Project Overview
+
+**What:** Ruby gem providing CLI productivity tools for AppyDave's YouTube content creation workflow.
+**Stack:** Ruby 3.4.2, Bundler 2.6.2, RSpec, RuboCop, semantic-release CI/CD.
+**This campaign:** 2 test additions + 1 production fix (1 line).
+**Commits:** Always use `kfix` — never `git commit`.
+
+---
+
+## Build & Run Commands
+
+```bash
+eval "$(rbenv init -)"
+
+RUBYOPT="-W0" bundle exec rspec                                                    # All tests
+bundle exec rspec spec/appydave/tools/jump/commands/add_spec.rb
+bundle exec rspec spec/appydave/tools/gpt_context/cli_spec.rb
+bundle exec rspec spec/appydave/tools/gpt_context/file_collector_spec.rb
+bundle exec rubocop --format clang
+```
+
+**Baseline:** 830 examples, 0 failures, 85.92% line coverage
+
+---
+
+## Directory Structure
+
+```
+lib/appydave/tools/gpt_context/
+  file_collector.rb         LINE 19 — change `return build_formats unless` to `return '' unless`
+
+spec/appydave/tools/jump/commands/
+  add_spec.rb               ADD `type` field to existing data integrity it block
+
+spec/appydave/tools/gpt_context/
+  cli_spec.rb               ADD new it block for `-f json` in the `-f format` describe block
+  file_collector_spec.rb    UPDATE nonexistent-dir test to confirm empty string (verify still passes)
+```
+
+---
+
+## Work Unit Details
+
+---
+
+### fix-b031 — add_spec: add type field to data integrity test
+
+**Read first:** `spec/appydave/tools/jump/commands/add_spec.rb`
+
+**Find the `'returns location data matching all input attrs'` it block** (added in final-test-gaps). It currently asserts: key, path, jump, tags, description.
+
+**Add one assertion** to that same it block:
+```ruby
+expect(location[:type]).to eq('tool')
+```
+
+`valid_attrs` has `type: 'tool'` so this will pass. The `location.to_h` in Location uses `.compact` — since type is set, it will be present.
+
+**Do NOT create a new it block** — add the assertion to the existing one.
+
+**Run:**
+```bash
+bundle exec rspec spec/appydave/tools/jump/commands/add_spec.rb
+```
+
+All 15 examples should pass. Then full suite. Then:
+```bash
+kfix "add type field assertion to add_spec location data integrity test"
+```
+
+---
+
+### fix-b032 — cli_spec: subprocess test for -f json
+
+**Read first:** `spec/appydave/tools/gpt_context/cli_spec.rb` — specifically the `-f format` describe block.
+
+**Add a new it block** to the `-f format` describe context:
+
+```ruby
+it 'outputs valid JSON when -f json specified' do
+  Dir.mktmpdir do |tmpdir|
+    File.write(File.join(tmpdir, 'test.rb'), '# test content')
+    outfile = File.join(tmpdir, 'output.txt')
+
+    `ruby #{script} -i '*.rb' -f json -b #{tmpdir} -o #{outfile} 2>&1`
+
+    content = File.read(outfile)
+    expect { JSON.parse(content) }.not_to raise_error
+    parsed = JSON.parse(content)
+    expect(parsed).to have_key('tree')
+    expect(parsed).to have_key('content')
+  end
+end
+```
+
+Note: `JSON` is available in the spec because Ruby stdlib is loaded. No require needed.
+
+**Run:**
+```bash
+bundle exec rspec spec/appydave/tools/gpt_context/cli_spec.rb
+```
+
+Then full suite. Then:
+```bash
+kfix "add subprocess test for -f json flag to cli_spec"
+```
+
+---
+
+### fix-b033 — file_collector.rb: fix silent CWD collection
+
+**Read first:** `lib/appydave/tools/gpt_context/file_collector.rb` line 19.
+
+Current code:
+```ruby
+return build_formats unless @working_directory && Dir.exist?(@working_directory)
+```
+
+**Change to:**
+```ruby
+return '' unless @working_directory && Dir.exist?(@working_directory)
+```
+
+This ensures a missing working directory always returns empty string rather than silently collecting files from the current process working directory.
+
+**Then verify the existing spec still passes:**
+```bash
+bundle exec rspec spec/appydave/tools/gpt_context/file_collector_spec.rb
+```
+
+The `'#build with nonexistent working directory'` example already asserts `expect(subject.build).to eq('')` — this should still pass (and now it's guaranteed by the code, not by an accident of the glob pattern).
+
+**Run rubocop:**
+```bash
+bundle exec rubocop lib/appydave/tools/gpt_context/file_collector.rb --format clang
+```
+
+Then full suite. Then:
+```bash
+kfix "fix file_collector silent collection from CWD when working directory does not exist"
+```
+
+---
+
+## Success Criteria
+
+- [ ] `RUBYOPT="-W0" bundle exec rspec` — 831+ examples, 0 failures
+- [ ] `bundle exec rubocop --format clang` — 0 offenses
+- [ ] Line coverage stays ≥ 85.92%
+- [ ] B031: `type` field asserted in add_spec data integrity test
+- [ ] B032: json format verified via subprocess in cli_spec
+- [ ] B033: `file_collector.rb:19` returns `''` not `build_formats` when dir missing
+
+---
+
+## Anti-Patterns to Avoid
+
+- ❌ Do NOT use `$?` — use `$CHILD_STATUS`
+- ❌ Do NOT use `git commit` — use `kfix`
+- ❌ Do NOT add `require 'spec_helper'`
+- ❌ Do NOT add multiple separate `before` blocks on same context (RSpec/ScatteredSetup)
+- ❌ Do NOT create new it blocks for B031 — add assertion to existing block
+
+---
+
+## Learnings (inherited from final-test-gaps)
+
+- **`$CHILD_STATUS` not `$?`** — rubocop Style/SpecialGlobalVars cop
+- **`options.format` defaults to `'tree,content'`** — never nil
+- **`options.prompt` defaults to nil** — aider format returns `''` when nil
+- **`location.to_h` uses symbol keys + `.compact`** — nil fields are dropped
+- **Agent pre-read pattern works** — read source files before writing assertions; zero wrong field name failures
+- **Parallel wave, all different files** — no merge conflicts expected
+- **JSON in subprocess test** — `JSON.parse(content)` in subprocess tests — ensure JSON output goes to outfile not stdout; use `-o outfile` flag

data/docs/planning/micro-cleanup/IMPLEMENTATION_PLAN.md

@@ -0,0 +1,27 @@
+# IMPLEMENTATION_PLAN.md — micro-cleanup
+
+**Goal**: Close 3 small items from the final-test-gaps quality audit: type field assertion (B031), -f json CLI test (B032), file_collector.rb silent-collection fix (B033)
+**Started**: 2026-03-19
+**Target**: All 3 complete; 830+ examples passing; rubocop clean; no regressions
+
+## Summary
+- Total: 3 | Complete: 0 | In Progress: 3 | Pending: 0 | Failed: 0
+
+## Pending
+
+## In Progress
+- [~] fix-b031 — add_spec: add `type` field assertion to location data integrity test
+- [~] fix-b032 — cli_spec: add subprocess test for `-f json` flag
+- [~] fix-b033 — file_collector.rb: return `''` directly when working_directory doesn't exist (line 19)
+
+## Complete
+
+## Failed / Needs Retry
+
+## Notes & Decisions
+- All 3 work units are independent — parallel wave
+- B033 is a production code change (1 line) — needs a spec update to confirm the behaviour
+- B031 and B032 are test-only
+- B033 fix: change `return build_formats unless` to `return '' unless` on line 19 of file_collector.rb
+- B031 fix: add `expect(location[:type]).to eq('tool')` to the existing 'returns location data matching all input attrs' it block
+- B032 fix: new it block in '-f format' describe context — subprocess with `-f json`, parse output as JSON, assert tree+content keys present

data/lib/appydave/tools/gpt_context/file_collector.rb

@@ -16,7 +16,7 @@ module Appydave
         end
 
         def build
-          return build_formats unless @working_directory && Dir.exist?(@working_directory)
+          return '' unless @working_directory && Dir.exist?(@working_directory)
 
           FileUtils.cd(@working_directory) { build_formats }
         end

data/lib/appydave/tools/version.rb

@@ -2,6 +2,6 @@
 
 module Appydave
   module Tools
-    VERSION = '0.76.3'
+    VERSION = '0.76.4'
   end
 end

data/package.json

@@ -1,6 +1,6 @@
 {
   "name": "appydave-tools",
-  "version": "0.76.3",
+  "version": "0.76.4",
   "description": "AppyDave YouTube Automation Tools",
   "scripts": {
     "release": "semantic-release"

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: appydave-tools
 version: !ruby/object:Gem::Version
-  version: 0.76.3
+  version: 0.76.4
 platform: ruby
 authors:
 - David Cruwys
@@ -302,10 +302,12 @@ files:
 - docs/planning/bugfix-and-security/assessment.md
 - docs/planning/final-test-gaps/AGENTS.md
 - docs/planning/final-test-gaps/IMPLEMENTATION_PLAN.md
+- docs/planning/final-test-gaps/assessment.md
 - docs/planning/fr2-gpt-context-help/AGENTS.md
 - docs/planning/fr2-gpt-context-help/IMPLEMENTATION_PLAN.md
 - docs/planning/fr2-gpt-context-help/assessment.md
-- docs/planning/next-round-brief.md
+- docs/planning/micro-cleanup/AGENTS.md
+- docs/planning/micro-cleanup/IMPLEMENTATION_PLAN.md
 - docs/planning/test-coverage-gaps/AGENTS.md
 - docs/planning/test-coverage-gaps/IMPLEMENTATION_PLAN.md
 - docs/planning/test-coverage-gaps/assessment.md

data/docs/planning/next-round-brief.md

@@ -1,42 +0,0 @@
-# Next Round Brief
-
-**Created:** 2026-03-19
-**Updated:** 2026-03-19 (after bugfix-and-security assessment)
-
----
-
-## Recommended Next Campaign: test-coverage-gaps
-
-### Goal
-
-Protect the B017 SSL security fix with a regression test, expand functional test coverage across gpt_context CLI and DAM range logic, and add the missing Jump Commands layer specs.
-
-### Background
-
-Quality audit after bugfix-and-security found:
-
-1. **B024** — `configure_ssl_options` has zero unit tests. The B017 SSL fix (removing unconditional `ssl_verify_peer: false`) has no regression protection. If reverted, all tests still pass. Must fix.
-2. **B022** — `cli_spec.rb` only tests `--help`, `--version`, no-args. No functional tests for `-i`, `-e`, `-f`, `-o`. Core behaviour untested at CLI level.
-3. **B026** — `determine_range` tests narrow (b40, b65, b99 only). Missing: b00, b9, a40. Both sync_from_ssd_spec and manifest_generator_spec need these.
-4. **B027** — gpt_context no-args spec only checks output string. Does not verify file collection stops.
-5. **B018** — Jump Commands (Remove/Add/Update) — zero dedicated specs.
-6. **B025** — Stale comment sync_from_ssd.rb line 173 (says 60-69, should say b50-b99).
-
-### Suggested Work Units (parallel — all test-only except B025)
-
-1. **fix-b024-ssl-tests** — Add `configure_ssl_options` unit tests to s3_operations_spec and share_operations_spec. Verify empty hash on default path; `{ssl_verify_peer: false}` when ENV override set. Stub ENV directly (`allow(ENV).to receive(:[]).with('AWS_SDK_RUBY_SKIP_SSL_VERIFICATION').and_return('true')`).
-2. **fix-b022-cli-tests** — Add functional subprocess tests to cli_spec.rb for -i, -e, -f, -o flags. Write to Tempfile, verify content. Use `Dir.mktmpdir` and clean up after.
-3. **fix-b026-b025-range-tests** — Add edge cases (b00, b9, a40) to sync_from_ssd_spec and manifest_generator_spec. Fix stale comment sync_from_ssd.rb line 173 while in the file.
-4. **fix-b027-noargs-test** — Strengthen no-args spec: `expect(Appydave::Tools::GptContext::FileCollector).not_to receive(:new)` when no patterns given.
-5. **fix-b018-jump-specs** — Add spec files for Jump Commands::Remove, Commands::Add, Commands::Update. Read existing Jump CLI spec first for setup pattern. Use JumpTestLocations + `with jump filesystem` context.
-
-### Mode Recommendation
-
-**Extend** — same stack, same patterns, test-only work. Inherit AGENTS.md.
-
-### Pre-Campaign Notes
-
-- Check if `climate_control` gem is in Gemfile before using ClimateControl — use direct ENV stubbing if not available
-- For B022 functional tests: subprocess writes to file, assert content includes `# file:` headers
-- For B027: stub at the class level, not instance — `expect(described_class).not_to receive(:new)`
-- For B018: read `spec/appydave/tools/jump/` existing specs before writing new command specs