@hustle-together/api-dev-tools 3.10.1 → 3.12.1

package/.claude/commands/worktree-add.md

@@ -0,0 +1,307 @@
+---
+description: Add a new git worktree from branch name or GitHub issue URL, copy settings, install deps, and open in current IDE
+argument-hint: <branch-name-or-github-issue-url> [optional-base-branch]
+---
+
+# Git Worktree Setup
+
+## General Guidelines
+
+### Output Style
+
+- **Never explicitly mention TDD** in code, comments, commits, PRs, or issues
+- Write natural, descriptive code without meta-commentary about the development process
+- The code should speak for itself - TDD is the process, not the product
+
+Create a new git worktree for branch: $ARGUMENTS
+
+<current_state>
+Current branch: !git branch --show-current`
+Current worktrees: !git worktree list`
+Remote branches: !git branch -r`
+Uncommitted changes: !git status --short`
+</current_state>
+
+<execution_steps>
+<step_0>
+<description>Validate MCP dependencies</description>
+<check_github_mcp>
+<requirement>GitHub MCP server must be configured</requirement>
+<fallback>If unavailable, use `gh` CLI commands</fallback>
+<validation> - Try listing available MCP resources - If GitHub MCP not found, switch to CLI fallback - Inform user about MCP configuration if needed
+</validation>
+</check_github_mcp>
+<error_handling>
+If MCP validation fails: - Show clear error message - Provide setup instructions - Fallback to CLI if possible
+</error_handling>
+<purpose>Ensure required MCP dependencies are available before proceeding</purpose>
+</step_0>
+
+<step_1>
+<description>Detect current IDE environment</description>
+<detection_methods>
+<method_1>
+<tool>mcp**ide**getDiagnostics</tool>
+<vs_code_insiders>Check for paths containing "Code - Insiders"</vs_code_insiders>
+<vs_code>Check for paths containing "Code/" or "Code.app"</vs_code>
+<cursor>Check for paths containing "Cursor"</cursor>
+<zed>Check for paths containing "Zed"</zed>
+</method_1>
+<method_2>
+<fallback_detection>Use which command to find available IDEs</fallback_detection>
+<commands>which code-insiders code zed cursor</commands>
+<priority_order>code-insiders > cursor > zed > code</priority_order>
+</method_2>
+</detection_methods>
+<set_variables>
+<ide_command>Detected command (code-insiders|code|zed|cursor)</ide_command>
+<ide_name>Human-readable name</ide_name>
+<supports_tasks>true for VS Code variants, false for others</supports_tasks>
+</set_variables>
+<purpose>Automatically detect which IDE to use for opening the new worktree</purpose>
+</step_1>
+
+<step_2>
+<description>Parse the arguments</description>
+<input>The user-provided arguments</input>
+<expected_format>branch-name-or-github-url [optional-base-branch]</expected_format>
+<example>fix/issue-123-main-content-area-visually-clipped main</example>
+<example_github_url><https://github.com/owner/project/issues/123> main</example_github_url>
+<default_base_branch>main (if not specified)</default_base_branch>
+</step_1>
+
+<step_2_5>
+<description>Handle GitHub issue URLs</description>
+<condition>If first argument matches GitHub issue URL pattern</condition>
+<url_detection>Check if argument contains "github.com" and "/issues/"</url_detection>
+<url_parsing>
+<pattern>https://github.com/{owner}/{repo}/issues/{issue_number}</pattern>
+<extract>owner, repo, issue_number from URL</extract>
+</url_parsing>
+<fetch_issue_details>
+<tool>mcp**github**issue_read</tool>
+<method>get</method>
+<parameters>owner, repo, issue_number</parameters>
+</fetch_issue_details>
+<generate_branch_name>
+<determine_type>Analyze issue title/labels to determine type (feat/fix/refactor/chore)</determine_type>
+<format>{type}/{repo}-{issue_number}-{kebab-case-title}</format>
+<kebab_case>Convert title to lowercase, replace spaces/special chars with hyphens</kebab_case>
+<sanitization>
+<rule>Always use lowercase for branch names</rule>
+<rule>Replace # with - (hash symbol not allowed in git branch names)</rule>
+<rule>Remove or replace other special characters: @, $, %, ^, &, \*, (, ), [, ], {, }, \, |, ;, :, ", ', <, >, ?, /, ~, `</rule>
+<rule>Replace multiple consecutive hyphens with single hyphen</rule>
+<rule>Trim leading/trailing hyphens</rule>
+</sanitization>
+<truncate>Limit total branch name to reasonable length (~60 chars)</truncate>
+</generate_branch_name>
+<user_confirmation>
+<display>Show generated branch name and ask for confirmation</display>
+<options>"Yes, proceed" or "No, exit" or "Edit branch name"</options>
+<if_no>Exit command gracefully</if_no>
+<if_edit>Allow user to modify the branch name</if_edit>
+<if_yes>Continue with generated/modified branch name</if_yes>
+</user_confirmation>
+<examples>
+<input>https://github.com/owner/project/issues/456</input>
+
+<title>"Fix duplicate items in list view"</title>
+<generated>fix/issue-456-duplicate-items-in-list-view</generated>
+</examples>
+</step_1_5>
+
+<step_3>
+<description>Add all files and stash uncommitted changes if any exist</description>
+<condition>If output is not empty (has uncommitted changes)</condition>
+<command>git add -A && git stash push -m "Worktree switch: Moving changes to ${branch_name}"</chained_command>
+<purpose>Preserve work in progress before switching worktrees</purpose>
+<note>Remember stash was created for later restoration</note>
+</step_3>
+
+<step_4>
+<description>Determine worktree parent directory</description>
+<check_if_in_worktree>git rev-parse --is-inside-work-tree && git worktree list --porcelain | grep "$(git rev-parse --show-toplevel)"</check_if_in_worktree>
+<set_parent_path>
+<if_main_worktree>Set parent_path=".."</if_main_worktree>
+<if_secondary_worktree>Set parent_path="../.." (need to go up two levels)</if_secondary_worktree>
+</set_parent_path>
+<purpose>Correctly determine where to create new worktree regardless of current location</purpose>
+<note>This handles both main worktree and secondary worktree scenarios</note>
+</step_4>
+
+<step_5>
+<description>Fetch latest changes from remote</description>
+<command>git fetch origin</command>
+<purpose>Ensure we have the latest remote branches and main branch state</purpose>
+<note>This ensures new worktrees are created from the most recent main branch</note>
+</step_6>
+
+<step_7>
+<description>Check if branch exists on remote</description>
+<command>git branch -r | grep "origin/${branch_name}"</command>
+<decision>
+<if_exists>Branch exists on remote - will checkout existing branch</if_exists>
+<if_not_exists>Branch does not exist - will create new branch from base</if_not_exists>
+</decision>
+</step_5>
+
+<step_6>
+<description>Create the worktree</description>
+<option_a_new_branch>
+<condition>Remote branch does NOT exist</condition>
+<command>git worktree add ${parent_path}/${branch_name} -b ${branch_name} ${base_branch}</command>
+      <example>git worktree add ../fix/issue-123-main-content-area-visually-clipped -b fix/issue-123-main-content-area-visually-clipped main</example>
+    </option_a_new_branch>
+    <option_b_existing_branch>
+      <condition>Remote branch EXISTS</condition>
+      <command>git worktree add ${parent_path}/${branch_name} ${branch_name}</command>
+<example>git worktree add ../fix/issue-123-main-content-area-visually-clipped fix/issue-123-main-content-area-visually-clipped</example>
+</option_b_existing_branch>
+</step_7>
+
+<step_8>
+<description>Copy Claude settings to new worktree</description>
+
+<source>.claude/settings.local.json</source>
+<destination>${parent_path}/${branch_name}/.claude/settings.local.json</destination>
+<command>cp -r .claude/settings.local.json ${parent_path}/${branch_name}/.claude/settings.local.json</command>
+<purpose>Preserve all permission settings and configurations</purpose>
+</step_8>
+
+<step*9>
+<description>Copy .env.local files to new worktree</description>
+<search_command>find . -name ".env.local" -type f</search_command>
+<copy_logic>For each .env.local file found, copy to corresponding location in new worktree</copy_logic>
+<common_locations> - app/.env.local - packages/\*/.env.local - (any other .env.local files found)
+</common_locations>
+<copy_command>find . -name ".env.local" -type f -exec sh -c 'mkdir -p "$(dirname "${parent_path}/${branch_name}/$1")" && cp "$1" "${parent_path}/${branch_name}/$1"' *{} \;</copy_command>
+<purpose>Preserve local environment configurations for development</purpose>
+<note>Only copies files that exist; ignores missing ones</note>
+</step_9>
+
+<step_10>
+<description>Create IDE-specific configuration (conditional)</description>
+<condition>Only if supports_tasks is true (VS Code variants)</condition>
+<vs_code_tasks>
+<create_directory>mkdir -p ${parent_path}/${branch_name}/.vscode</create_directory>
+<create_file_command>cat > ${parent_path}/${branch_name}/.vscode/tasks.json << 'EOF'
+{
+"version": "2.0.0",
+"tasks": [
+{
+"label": "Auto start Claude",
+"type": "shell",
+"command": "claude",
+"runOptions": {
+"runOn": "folderOpen"
+},
+"presentation": {
+"echo": false,
+"reveal": "always",
+"focus": true,
+"panel": "new"
+}
+}
+]
+}
+EOF</create_file_command>
+</vs_code_tasks>
+<purpose>Create auto-start Claude task for VS Code variants on folder open</purpose>
+<note>Only creates for VS Code variants (code, code-insiders, cursor)</note>
+<skip_message>Skipping IDE-specific config for non-VS Code IDEs</skip_message>
+</step_10>
+
+<step_11>
+<description>Install dependencies in new worktree</description>
+<working_directory>${parent_path}/${branch_name}</working_directory>
+<command>cd ${parent_path}/${branch_name} && pnpm install</command>
+<purpose>Ensure all node_modules are installed for the new worktree</purpose>
+</step_11>
+
+<step_12>
+<description>Apply stashed changes to new worktree (if stash was created)</description>
+<condition>Only if stash was created in step_3</condition>
+<working_directory>${parent_path}/${branch_name}</working_directory>
+<command>cd ${parent_path}/${branch_name} && git stash pop</command>
+<purpose>Restore uncommitted work-in-progress to the new worktree branch</purpose>
+<note>This moves your uncommitted changes to the new branch where you'll continue working</note>
+</step_12>
+
+<step_13>
+<description>Open detected IDE in new worktree</description>
+<command>${ide_command} ${parent_path}/${branch_name}</command>
+<ide_specific_behavior>
+<vs_code_variants>Opens folder in VS Code/Insiders/Cursor with tasks.json auto-starting Claude</vs_code_variants>
+<zed>Opens folder in Zed editor</zed>
+<other>Uses detected IDE command to open folder</other>
+</ide_specific_behavior>
+<purpose>Launch development environment for the new worktree using detected IDE</purpose>
+<confirmation_message>Opening worktree in ${ide_name}</confirmation_message>
+</step_11>
+</execution_steps>
+
+<important_notes>
+
+- Automatically detects and uses your current IDE (VS Code, VS Code Insiders, Cursor, Zed, etc.)
+- Creates VS Code-specific tasks.json only for VS Code variants (auto-starts Claude on folder open)
+- Branch names with slashes (feat/, fix/, etc.) are fully supported
+- The worktree directory path will match the full branch name including slashes
+- Settings are copied to maintain the same permissions across worktrees
+- Environment files (.env.local) are copied to preserve local configurations
+- Each worktree has its own node_modules installation
+- Uncommitted changes are automatically stashed and moved to the new worktree
+- Your work-in-progress seamlessly transfers to the new branch
+- IDE detection fallback: checks available editors and uses priority order
+  </important_notes>
+
+## 🛡 Project Rules (Injected into every command)
+
+1. **NO BROKEN BUILDS:**
+   - Run `pnpm test` before every `/commit`
+   - Ensure all tests pass
+   - Fix any type errors immediately
+
+2. **API DEVELOPMENT:**
+   - All new APIs MUST have Zod request/response schemas
+   - All APIs MUST be documented in both:
+     - OpenAPI spec ([src/lib/openapi/](src/lib/openapi/))
+     - API test manifest ([src/app/api-test/api-tests-manifest.json](src/app/api-test/api-tests-manifest.json))
+   - Test ALL parameters and edge cases
+   - Include code examples and real-world outputs
+
+3. **TDD WORKFLOW:**
+   - ALWAYS use /red → /green → /refactor cycle
+   - NEVER write implementation without failing test first
+   - Use /cycle for feature development
+   - Use characterization tests for refactoring
+
+4. **API KEY MANAGEMENT:**
+   - Support three loading methods:
+     - Server environment variables
+     - NEXT*PUBLIC* variables (client-side)
+     - Custom headers (X-OpenAI-Key, X-Anthropic-Key, etc.)
+   - Never hardcode API keys
+   - Always validate key availability before use
+
+5. **COMPREHENSIVE TESTING:**
+   - When researching APIs, read actual implementation code
+   - Discover ALL possible parameters (not just documented ones)
+   - Test with various parameter combinations
+   - Document custom headers, query params, request/response schemas
+   - Include validation rules and testing notes
+
+6. **NO UI BLOAT:**
+   - This is an API project with minimal frontend
+   - Only keep necessary test/documentation interfaces
+   - Delete unused components immediately
+   - No unnecessary UI libraries or features
+
+7. **DOCUMENTATION:**
+   - If you change an API, you MUST update:
+     - OpenAPI spec
+     - api-tests-manifest.json
+     - Code examples
+     - Testing notes
+   - Document expected behavior and edge cases
+   - Include real-world output examples

package/.claude/commands/worktree-cleanup.md

@@ -0,0 +1,275 @@
+---
+description: Clean up merged worktrees by verifying PR/issue status, consolidating settings, and removing stale worktrees
+argument-hint: (no arguments)
+---
+
+# Worktree Cleanup
+
+## General Guidelines
+
+### Output Style
+
+- **Never explicitly mention TDD** in code, comments, commits, PRs, or issues
+- Write natural, descriptive code without meta-commentary about the development process
+- The code should speak for itself - TDD is the process, not the product
+
+Clean up merged worktrees by finding the oldest merged branch, consolidating settings, and removing stale worktrees.
+
+Additional info: $ARGUMENTS
+
+<current_state>
+Current branch: !git branch --show-current`
+Current worktrees: !git worktree list`
+</current_state>
+
+<execution_steps>
+<step_0>
+<description>Validate MCP dependencies</description>
+<check_github_mcp>
+<requirement>GitHub MCP server must be configured</requirement>
+<fallback>If unavailable, use `gh` CLI commands</fallback>
+<validation> - Try listing available MCP resources - If GitHub MCP not found, switch to CLI fallback - Inform user about MCP configuration if needed
+</validation>
+</check_github_mcp>
+<error_handling>
+If MCP validation fails: - Show clear error message - Provide setup instructions - Fallback to CLI if possible
+</error_handling>
+<purpose>Ensure required MCP dependencies are available before proceeding</purpose>
+</step_0>
+
+<step_1>
+<description>Verify we're in main branch</description>
+<check_command>git branch --show-current</check_command>
+<required_branch>main</required_branch>
+<error_if_not_main>Exit with error message: "This command must be run from the main branch"</error_if_not_main>
+<purpose>Ensure we're consolidating to the main worktree</purpose>
+</step_1>
+
+<step_2>
+<description>Get list of all worktrees</description>
+<command>git worktree list --porcelain</command>
+<parse_output>Extract worktree paths and branch names</parse_output>
+<exclude_main>Filter out the main worktree from cleanup candidates</exclude_main>
+<purpose>Identify all worktrees that could potentially be cleaned up</purpose>
+</step_2>
+
+<step_3>
+<description>Find oldest worktree by directory age</description>
+<get_worktree_ages>
+<command_macos>git worktree list | grep -v "main" | awk '{print $1}' | while read path; do /usr/bin/stat -f "%Sm|%N" -t "%Y-%m-%d %H:%M:%S" "$path" 2>/dev/null; done | sort</command_macos>
+<command_linux>git worktree list | grep -v "main" | awk '{print $1}' | xargs stat -c "%y|%n" | sort</command_linux>
+<purpose>List all worktrees sorted by directory modification time (oldest first)</purpose>
+<note>Use full path /usr/bin/stat on macOS, regular stat on Linux.</note>
+</get_worktree_ages>
+<filter_recent>
+<exclude_new>For worktrees created within the last 24 hours, let user know that this worktree might not be worth cleaning</exclude_new>
+<get_current_time>date +"%Y-%m-%d %H:%M"</get_current_time>
+</filter_recent>
+<select_oldest>
+<extract_branch_name>Parse branch name from oldest worktree path</extract_branch_name>
+<important_note>DO NOT use "git branch --merged" to check merge status - it's unreliable</important_note>
+<proceed_to_pr_check>Move directly to step 4 to verify PR merge status instead</proceed_to_pr_check>
+</select_oldest>
+<purpose>Identify oldest worktree candidate - actual merge verification happens via GitHub PR in next step</purpose>
+</step_3>
+
+<step_4>
+<description>Verify GitHub PR merge status (primary merge verification)</description>
+<determine_repo>
+<check_remote>git remote get-url origin</check_remote>
+<parse_repo>Extract owner/repo from GitHub URL</parse_repo>
+<fallback>Use project repository from git remote (owner/repo format)</fallback>
+</determine_repo>
+<search_pr>
+<tool>mcp**github**search_pull_requests</tool>
+<query>repo:owner/repo head:{branch_name} base:main</query>
+<purpose>Find PR for this branch targeting main</purpose>
+<important>This is the PRIMARY way to verify if a branch was merged - NOT git commands</important>
+</search_pr>
+<verify_pr_merged>
+<if_pr_found>
+<get_pr_details>Use mcp**github**pull_request_read to get full PR info</get_pr_details>
+<confirm_merged>Verify PR state is "closed" AND merged_at is not null AND base is "main"</confirm_merged>
+<extract_issue_number>Look for issue references in PR title/body (e.g., #14533, owner/repo#14533)</extract_issue_number>
+<if_merged>Proceed with cleanup - this branch was definitively merged to main</if_merged>
+<if_not_merged>
+<skip_worktree>This worktree is NOT merged - continue to next oldest worktree</skip_worktree>
+<repeat_from_step_3>Go back and find the next oldest worktree to check</repeat_from_step_3>
+</if_not_merged>
+</if_pr_found>
+<if_no_pr>
+<skip_worktree>No PR found - this branch was likely never submitted for review</skip_worktree>
+<continue_to_next>Continue checking next oldest worktree</continue_to_next>
+</if_no_pr>
+</verify_pr_merged>
+<purpose>Use GitHub PR status as the authoritative source for merge verification instead of unreliable git commands</purpose>
+</step_4>
+
+<step_4_5>
+<description>Check and close related GitHub issue</description>
+<if_issue_found>
+<get_issue_details>
+<tool>mcp**github**issue_read</tool>
+<method>get</method>
+<extract_repo>From issue reference (main-repo vs cross-repo)</extract_repo>
+</get_issue_details>
+<check_issue_state>
+<if_open>
+<ask_close>Ask user: "Related issue #{number} is still open. Should I close it? (y/N)"</ask_close>
+<if_yes_close>
+<add_closing_comment>
+<tool>mcp**github**add_issue_comment</tool>
+<body_template>Closing this issue as branch {branch_name} was merged to main on {merge_date} via PR #{pr_number}.</body_template>
+<get_merge_date>Extract merge date from PR details</get_merge_date>
+<get_pr_number>Use PR number from search results</get_pr_number>
+</add_closing_comment>
+<close_issue>
+<tool>mcp**github**issue_write</tool>
+<method>update</method>
+<state>closed</state>
+<state_reason>completed</state_reason>
+</close_issue>
+</if_yes_close>
+</if_open>
+<if_closed>Inform user issue is already closed</if_closed>
+</check_issue_state>
+</if_issue_found>
+<if_no_issue>Continue without issue management</if_no_issue>
+<purpose>Ensure proper issue lifecycle management</purpose>
+</step_4_5>
+
+<step_5>
+<description>Check if worktree is locked</description>
+<check_command>git worktree list --porcelain | grep -A5 "worktree {path}" | grep "locked"</check_command>
+<if_locked>
+<unlock_command>git worktree unlock {path}</unlock_command>
+<notify_user>Inform user that worktree was unlocked</notify_user>
+</if_locked>
+<purpose>Unlock worktree if it was locked for tracking purposes</purpose>
+</step_5>
+
+<step_6>
+<description>Analyze Claude settings differences</description>
+<read_main_settings>.claude/settings.local.json</read_main_settings>
+<read_worktree_settings>{worktree_path}/.claude/settings.local.json</read_worktree_settings>
+<compare_allow_lists>
+<extract_main_allows>Extract "allow" array from main settings</extract_main_allows>
+<extract_worktree_allows>Extract "allow" array from worktree settings</extract_worktree_allows>
+<find_differences>Identify entries in worktree that are not in main</find_differences>
+</compare_allow_lists>
+<filter_suggestions>
+<include_filesystem>Read permissions for filesystem paths</include_filesystem>
+<exclude_intrusive>Exclude bash commands, write permissions, etc.</exclude_intrusive>
+<focus_user_specific>Include only user-specific, non-disruptive entries</focus_user_specific>
+</filter_suggestions>
+<purpose>Identify useful settings to consolidate before cleanup</purpose>
+</step_6>
+
+<step_7>
+<description>Suggest settings consolidation</description>
+<if_differences_found>
+<display_suggestions>Show filtered differences to user</display_suggestions>
+<ask_confirmation>Ask user which entries to add to main settings</ask_confirmation>
+<apply_changes>Update main .claude/settings.local.json with selected entries</apply_changes>
+</if_differences_found>
+<if_no_differences>Inform user no settings need consolidation</if_no_differences>
+<purpose>Preserve useful development settings before removing worktree</purpose>
+</step_7>
+
+<step_8>
+<description>Final cleanup confirmation</description>
+
+<summary>
+<display_worktree>Show worktree path and branch name</display_worktree>
+<show_pr_status>Show merged PR details if found</show_pr_status>
+<show_issue_status>Show related issue status if found</show_issue_status>
+<show_last_activity>Display directory creation/modification date</show_last_activity>
+</summary>
+<safety_checks>
+<check_uncommitted>git status --porcelain in worktree directory</check_uncommitted>
+<warn_if_dirty>Alert user if uncommitted changes exist</warn_if_dirty>
+</safety_checks>
+<ask_deletion>Ask user confirmation: "Delete this worktree? (y/N)"</ask_deletion>
+<purpose>Final safety check before irreversible deletion</purpose>
+</step_8>
+
+<step_9>
+<description>Delete worktree</description>
+<if_confirmed>
+<remove_worktree>git worktree remove {path} --force</remove_worktree>
+<cleanup_branch>git branch -d {branch_name}</cleanup_branch>
+<success_message>Inform user worktree was successfully removed</success_message>
+<next_steps>Suggest running command again to find next candidate</next_steps>
+</if_confirmed>
+<if_declined>Exit gracefully with no changes</if_declined>
+<purpose>Perform the actual cleanup and guide user for next iteration</purpose>
+</step_9>
+</execution_steps>
+
+<important_notes>
+
+- Uses GitHub PR merge status as the ONLY reliable way to verify if a branch was merged
+- DOES NOT use "git branch --merged" command as it's unreliable for merge verification
+- Only processes branches with PRs that were definitively merged to main
+- Skips worktrees without merged PRs and continues to next oldest candidate
+- Checks and optionally closes related GitHub issues
+- Prioritizes oldest worktrees by directory age first for systematic cleanup
+- Warns about very recent worktrees (created within 24 hours) to avoid cleaning active work
+- Preserves useful development settings before deletion
+- Requires explicit confirmation before any destructive actions
+- Handles locked worktrees automatically
+- Processes one worktree at a time to maintain control
+- Must be run from main branch for safety
+- Works with standard GitHub repository URLs (owner/repo format)
+  </important_notes>
+
+## 🛡 Project Rules (Injected into every command)
+
+1. **NO BROKEN BUILDS:**
+   - Run `pnpm test` before every `/commit`
+   - Ensure all tests pass
+   - Fix any type errors immediately
+
+2. **API DEVELOPMENT:**
+   - All new APIs MUST have Zod request/response schemas
+   - All APIs MUST be documented in both:
+     - OpenAPI spec ([src/lib/openapi/](src/lib/openapi/))
+     - API test manifest ([src/app/api-test/api-tests-manifest.json](src/app/api-test/api-tests-manifest.json))
+   - Test ALL parameters and edge cases
+   - Include code examples and real-world outputs
+
+3. **TDD WORKFLOW:**
+   - ALWAYS use /red → /green → /refactor cycle
+   - NEVER write implementation without failing test first
+   - Use /cycle for feature development
+   - Use characterization tests for refactoring
+
+4. **API KEY MANAGEMENT:**
+   - Support three loading methods:
+     - Server environment variables
+     - NEXT*PUBLIC* variables (client-side)
+     - Custom headers (X-OpenAI-Key, X-Anthropic-Key, etc.)
+   - Never hardcode API keys
+   - Always validate key availability before use
+
+5. **COMPREHENSIVE TESTING:**
+   - When researching APIs, read actual implementation code
+   - Discover ALL possible parameters (not just documented ones)
+   - Test with various parameter combinations
+   - Document custom headers, query params, request/response schemas
+   - Include validation rules and testing notes
+
+6. **NO UI BLOAT:**
+   - This is an API project with minimal frontend
+   - Only keep necessary test/documentation interfaces
+   - Delete unused components immediately
+   - No unnecessary UI libraries or features
+
+7. **DOCUMENTATION:**
+   - If you change an API, you MUST update:
+     - OpenAPI spec
+     - api-tests-manifest.json
+     - Code examples
+     - Testing notes
+   - Document expected behavior and edge cases
+   - Include real-world output examples