Adding prompt tool with first 9 prompts for onboarding#233
Conversation
|
Caution Review failedThe pull request is closed. WalkthroughAdds an onboarding prompts feature: new prompts data, a prompts tool and schema, server integration and telemetry for prompt actions, onboarding state and usage tracking, build changes to include data in dist, documentation updates, .gitignore update, and removal of a binary-file detection test. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
actor User
participant Client
participant Server
participant Prompts as "tools/prompts.ts"
participant Usage as "usageTracker"
participant Data as "src/data/onboarding-prompts.json"
User->>Client: Request prompts (list/get)
Client->>Server: CallToolRequest(name:get_prompts, args)
Server->>Prompts: getPrompts(args)
Prompts->>Data: loadPromptsData()
Data-->>Prompts: PromptsData
alt action == get_prompt
Prompts->>Usage: markPromptUsed(promptId, category)
Prompts-->>Server: Prompt response
Server->>Server: emit server_get_prompt telemetry
else list_categories / list_prompts
Prompts-->>Server: List/Category response
Server->>Server: emit list/category telemetry
end
Server-->>Client: Tool result
opt After successful call
Server->>Usage: shouldShowOnboarding()
alt true
Server->>Usage: getOnboardingMessage()
Server->>Usage: markOnboardingShown(variant)
Server-->>Client: Inject onboarding message into LLM response
end
opt onboarding prompts used
Server->>Usage: markOnboardingPromptsUsed()
end
end
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Possibly related PRs
Suggested labels
Suggested reviewers
Poem
Tip 👮 Agentic pre-merge checks are now available in preview!Pro plan users can now enable pre-merge checks in their settings to enforce checklists before merging PRs.
Please see the documentation for more information. Example: reviews:
pre_merge_checks:
custom_checks:
- name: "Undocumented Breaking Changes"
mode: "warning"
instructions: |
Pass/fail criteria: All breaking changes to public APIs, CLI flags, environment variables, configuration keys, database schemas, or HTTP/GraphQL endpoints must be documented in the "Breaking Change" section of the PR description and in CHANGELOG.md. Exclude purely internal or private changes (e.g., code not exported from package entry points or explicitly marked as internal).Please share your feedback with us on this Discord post. 📜 Recent review detailsConfiguration used: CodeRabbit UI Review profile: CHILL Plan: Pro 📒 Files selected for processing (2)
✨ Finishing Touches
🧪 Generate unit tests
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 5
🧹 Nitpick comments (5)
package.json (1)
30-30: Build step LGTM; consider copying all data JSONs.If more prompt catalogs are added, copy them generically to avoid future script edits.
- "build": "tsc && shx cp setup-claude-server.js uninstall-claude-server.js track-installation.js dist/ && shx chmod +x dist/*.js && shx mkdir -p dist/data && shx cp src/data/onboarding-prompts.json dist/data/", + "build": "tsc && shx cp setup-claude-server.js uninstall-claude-server.js track-installation.js dist/ && shx chmod +x dist/*.js && shx mkdir -p dist/data && shx cp src/data/*.json dist/data/"src/utils/usageTracker.ts (1)
356-363: Replace console log with structured analytics event.Leverage existing capture() so usage is observable and queryable.
async markPromptUsed(promptId: string, category: string): Promise<void> { - // This could be expanded later to track detailed prompt usage - // For now, we'll just rely on the capture analytics - console.log(`[PROMPT USAGE] User retrieved prompt: ${promptId} (category: ${category})`); + await capture('prompt_used', { + prompt_id: promptId, + category, + }); }src/tools/schemas.ts (1)
124-131: Make args validation stricter with a discriminated union.Enforces promptId only for get_prompt at the schema level.
-export const ListSearchesArgsSchema = z.object({}); +export const ListSearchesArgsSchema = z.object({}); -// Prompts tool schema -export const GetPromptsArgsSchema = z.object({ - action: z.enum(['list_categories', 'list_prompts', 'get_prompt']), - category: z.string().optional(), - promptId: z.string().optional(), -}); +// Prompts tool schema +export const GetPromptsArgsSchema = z.discriminatedUnion('action', [ + z.object({ action: z.literal('list_categories') }), + z.object({ + action: z.literal('list_prompts'), + category: z.string().optional(), + }), + z.object({ + action: z.literal('get_prompt'), + promptId: z.string(), + }), +]);src/server.ts (1)
681-710: Consider exposing prompts via MCP prompts capability too.Keeping get_prompts is fine; additionally wiring real prompts into ListPromptsRequest could let clients render them natively.
src/tools/prompts.ts (1)
321-335: Tone nit: don’t promise auto-execution.If the tool doesn’t actually auto-run, soften the claim to avoid user confusion.
- response += `## Ready to Use This Prompt\nThe prompt below is ready to use. I'll start executing it right away:\n\n`; + response += `## Ready to Use This Prompt\nThe prompt below is ready to use. I can execute it on your confirmation:\n\n`;
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
💡 Knowledge Base configuration:
- MCP integration is disabled by default for public repositories
- Jira integration is disabled by default for public repositories
- Linear integration is disabled by default for public repositories
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (6)
package.json(1 hunks)src/data/onboarding-prompts.json(1 hunks)src/server.ts(3 hunks)src/tools/prompts.ts(1 hunks)src/tools/schemas.ts(1 hunks)src/utils/usageTracker.ts(1 hunks)
🧰 Additional context used
🧬 Code graph analysis (2)
src/tools/prompts.ts (3)
src/types.ts (1)
ServerResult(51-55)src/utils/capture.ts (1)
capture(277-284)src/utils/usageTracker.ts (1)
usageTracker(367-367)
src/server.ts (2)
src/tools/schemas.ts (1)
GetPromptsArgsSchema(127-131)src/tools/prompts.ts (1)
getPrompts(65-129)
🔇 Additional comments (7)
src/data/onboarding-prompts.json (1)
1-114: No issues detected—prompt IDs are unique and all prompts include the required fields.src/server.ts (3)
46-47: Import looks good.
51-51: Prompts handler import LGTM.
777-788: Execution path LGTM with consistent error handling.src/tools/prompts.ts (3)
154-158: Confirm renderer interprets Markdown for type "text".Responses are Markdown-formatted but use
{ type: "text" }. Verify the client renders Markdown for this type; otherwise switch to the appropriate content type.Also applies to: 188-193, 229-233
260-276: LGTM on categories presentation.Clear summary with counts and guidance.
48-49: Data file packaging verified. The build script in package.json copiessrc/data/onboarding-prompts.jsonintodist/data/, matching the runtime path (dist/data/onboarding-prompts.json); no changes needed.
| if (!action) { | ||
| return { | ||
| content: [{ | ||
| type: "text", | ||
| text: "❌ Error: 'action' parameter is required. Use 'list_categories', 'list_prompts', or 'get_prompt'" | ||
| }], | ||
| isError: true | ||
| }; | ||
| } |
There was a problem hiding this comment.
🛠️ Refactor suggestion
Validate action value before emitting analytics.
Guard against arbitrary action strings to keep analytics clean and return early on invalid input.
if (!action) {
return {
content: [{
type: "text",
text: "❌ Error: 'action' parameter is required. Use 'list_categories', 'list_prompts', or 'get_prompt'"
}],
isError: true
};
}
+
+ const validActions = new Set(['list_categories','list_prompts','get_prompt']);
+ if (!validActions.has(action as any)) {
+ return {
+ content: [{ type: "text", text: "❌ Error: Invalid action. Use 'list_categories', 'list_prompts', or 'get_prompt'" }],
+ isError: true
+ };
+ }📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| if (!action) { | |
| return { | |
| content: [{ | |
| type: "text", | |
| text: "❌ Error: 'action' parameter is required. Use 'list_categories', 'list_prompts', or 'get_prompt'" | |
| }], | |
| isError: true | |
| }; | |
| } | |
| if (!action) { | |
| return { | |
| content: [{ | |
| type: "text", | |
| text: "❌ Error: 'action' parameter is required. Use 'list_categories', 'list_prompts', or 'get_prompt'" | |
| }], | |
| isError: true | |
| }; | |
| } | |
| const validActions = new Set(['list_categories', 'list_prompts', 'get_prompt']); | |
| if (!validActions.has(action as any)) { | |
| return { | |
| content: [{ | |
| type: "text", | |
| text: "❌ Error: Invalid action. Use 'list_categories', 'list_prompts', or 'get_prompt'" | |
| }], | |
| isError: true | |
| }; | |
| } |
🤖 Prompt for AI Agents
In src/tools/prompts.ts around lines 70-78, the current code only checks for
missing action but allows any string and may emit analytics for invalid values;
add a whitelist check against allowed actions
['list_categories','list_prompts','get_prompt'] immediately after the existing
null/undefined check and return the same error response (isError: true with the
message) if the action is not one of those values, ensuring no analytics are
emitted for invalid actions.
| // Track analytics for tool usage | ||
| await capture(`prompts_tool_${action}`, { | ||
| category: category, | ||
| prompt_id: promptId, | ||
| has_category_filter: !!category | ||
| }); | ||
|
|
There was a problem hiding this comment.
Don’t let telemetry or usage tracking block or fail the tool.
Make analytics fire-and-forget and swallow errors, especially in the catch block.
- await capture(`prompts_tool_${action}`, {
+ void capture(`prompts_tool_${action}`, {
category: category,
prompt_id: promptId,
has_category_filter: !!category
- });
+ }).catch(() => {});- await capture('prompts_tool_error', {
+ void capture('prompts_tool_error', {
error_message: error instanceof Error ? error.message : String(error),
action: params?.action
- });
+ }).catch(() => {});- await capture('prompt_retrieved', {
+ void capture('prompt_retrieved', {
prompt_id: promptId,
prompt_title: prompt.title,
category: prompt.categories[0] || 'uncategorized',
author: prompt.author,
verified: prompt.verified
- });
+ }).catch(() => {});
- await usageTracker.markPromptUsed(promptId, prompt.categories[0] || 'uncategorized');
+ void usageTracker.markPromptUsed(promptId, prompt.categories[0] || 'uncategorized').catch(() => {});Also applies to: 115-120, 214-225
🤖 Prompt for AI Agents
In src/tools/prompts.ts around lines 80-86 (and also update the similar calls at
115-120 and 214-225): the telemetry capture calls currently use await and can
block or propagate failures; change them to fire-and-forget by removing the
await and appending a rejection handler (e.g., call capture(...).catch(()=>{})
or use void capture(...).catch(()=>{})) so errors are swallowed and telemetry
cannot block or fail the tool.
| async function listPrompts(category?: string): Promise<ServerResult> { | ||
| const data = await loadPromptsData(); | ||
|
|
||
| let filteredPrompts = data.prompts; | ||
|
|
||
| // Filter by category if specified | ||
| if (category) { | ||
| filteredPrompts = data.prompts.filter(prompt => | ||
| prompt.categories.includes(category) | ||
| ); | ||
|
|
||
| if (filteredPrompts.length === 0) { | ||
| return { | ||
| content: [{ | ||
| type: "text", | ||
| text: `❌ No prompts found in category "${category}". Use action='list_categories' to see available categories.` | ||
| }], | ||
| isError: true | ||
| }; | ||
| } | ||
| } |
There was a problem hiding this comment.
🛠️ Refactor suggestion
Make category filtering tolerant (trim + case-insensitive).
Prevents “No prompts found” due to whitespace/case mismatches.
- let filteredPrompts = data.prompts;
+ let filteredPrompts = data.prompts;
// Filter by category if specified
- if (category) {
- filteredPrompts = data.prompts.filter(prompt =>
- prompt.categories.includes(category)
- );
+ const normalizedCategory = category?.trim().toLowerCase();
+ if (normalizedCategory) {
+ filteredPrompts = data.prompts.filter(prompt =>
+ prompt.categories.some(c => c.toLowerCase() === normalizedCategory)
+ );Committable suggestion skipped: line range outside the PR's diff.
🤖 Prompt for AI Agents
In src/tools/prompts.ts around lines 164 to 184, the category filter is strict
and can miss matches due to whitespace or casing; modify it to normalize both
the incoming category and each prompt's categories by trimming and lowercasing
before comparing (e.g., compute const normalizedCategory =
category.trim().toLowerCase() then filter prompts where prompt.categories.some(c
=> c.trim().toLowerCase() === normalizedCategory)); keep the early return
behavior when no matches but ensure you only call .trim()/.toLowerCase() after
confirming category is non-empty.
- Add OnboardingState interface to usageTracker - Implement shouldShowOnboarding() logic (triggers after 3-10 successful calls) - Add onboarding message injection to server.ts (before feedback prompts) - Track onboarding analytics: message_shown, message_accepted - Auto-detect onboarding acceptance when user calls get_prompts - Use Variant A: Welcome & Guidance message for first-time users - Complete analytics pipeline for cohort analysis and A/B testing
…for accuracy - Changed onboarding_message_accepted to user_asked_for_prompts - Renamed onboardingAccepted to onboardingUsedPrompts in state - Updated method names to be more accurate - User may call get_prompts for reasons other than responding to onboarding invitation
- get_prompts calls are already captured by standard tool call analytics - prompts_tool_* events provide specific action tracking - Simplified to just update state for onboarding flow control - Eliminates duplicate/redundant analytics data
- Removed redundant prompts_tool_* events from prompts.ts - Added get_prompts parameters (action, category, prompt_id) to server_call_tool event - Added detailed prompt_retrieved event for specific prompt usage - Cleaner analytics with all prompts data in standard tool call events - Exported loadPromptsData function for server analytics use
- Changed onboarding_message_shown -> server_onboarding_shown - Changed prompt_retrieved -> server_prompt_retrieved - Follows established naming convention: - package_* for npm/installation events - server_* for server-side operations - Consistent with server_call_tool, server_request_error patterns
There was a problem hiding this comment.
Actionable comments posted: 3
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
README.md (1)
102-105: Auto-Updates text is self-contradictory“✅ Auto-Updates: Yes - requires manual updates” is unclear. If Option 2 (bash) truly auto-updates on restart, drop “requires manual updates”; if not, flip to “No”.
Apply one of these diffs:
-**✅ Auto-Updates:** Yes - requires manual updates +**✅ Auto-Updates:** Yesor
-**✅ Auto-Updates:** Yes - requires manual updates +**❌ Auto-Updates:** No - re-run the installer to update
♻️ Duplicate comments (2)
src/tools/prompts.ts (2)
170-184: Make category filtering tolerant (trim + case-insensitive)Prevents “No prompts found” due to whitespace/case mismatches.
- if (category) { - filteredPrompts = data.prompts.filter(prompt => - prompt.categories.includes(category) - ); + const normalizedCategory = category?.trim().toLowerCase(); + if (normalizedCategory) { + filteredPrompts = data.prompts.filter(prompt => + prompt.categories.some(c => c.toLowerCase() === normalizedCategory) + );
56-66: Harden JSON parsing with shape validation (truthy check is ineffective)Validate the parsed structure before caching to prevent malformed data from slipping through.
try { const dataPath = path.join(__dirname, '..', 'data', 'onboarding-prompts.json'); const fileContent = await fs.readFile(dataPath, 'utf-8'); - cachedPromptsData = JSON.parse(fileContent); - - if (!cachedPromptsData) { - throw new Error('Failed to parse prompts data'); - } + const parsed = JSON.parse(fileContent); + if (!parsed || typeof parsed !== 'object' || !Array.isArray((parsed as any).prompts)) { + throw new Error('Invalid prompts data: expected { prompts: Prompt[] }'); + } + cachedPromptsData = parsed as PromptsData;
🧹 Nitpick comments (13)
FAQ.md (1)
176-193: Onboarding guidance is concise and alignedReads well and matches the described flow (auto + manual + examples). Consider adding a one-liner hint that these are powered by the new get_prompts tool to connect docs to UX, but optional.
README.md (1)
147-147: Typo: duplicated heading markers
### ### Option 5→### Option 5.-### ### Option 5: Checkout locally ❌ **Manual Updates** **Requires Node.js** ❌ **Manual Updates** **Requires Node.js** +### Option 5: Checkout locally ❌ **Manual Updates** **Requires Node.js** ❌ **Manual Updates** **Requires Node.js**test_onboarding.md (2)
37-43: Markdown lint: add a language to this fenced blockLabel the debug output block to satisfy MD040 and improve rendering.
-``` +```text [ONBOARDING DEBUG] Should show onboarding: true ...--- `64-81`: **Production/test delay toggling should be config-driven** Instead of editing code to switch delays, consider reading delays from config or env (e.g., ONBOARDING_DELAY_MS list) so tests can override without code changes. </blockquote></details> <details> <summary>.serena/project.yml (1)</summary><blockquote> `25-25`: **Trailing space in comment** YAML linters flag this. Remove trailing spaces on Line 25. ```diff -# To make sure you have the latest list of tools, and to view their descriptions, +# To make sure you have the latest list of tools, and to view their descriptions,src/data/onboarding-prompts.json (1)
1-4: Add a top-level “lastUpdated” field (optional)A timestamp helps consumers handle cache invalidation/versioning without bumping semantic version for content-only tweaks.
src/server.ts (2)
747-747: Make telemetry fire-and-forget to avoid latency and unhandled rejectionsCurrent capture calls can block the request path or generate unhandled promise rejections. Prefer non-blocking calls with explicit rejection swallowing.
Apply diffs (representative spots; repeat the pattern for other capture(...) calls in this file):
- capture_call_tool('server_call_tool', telemetryData); + void capture_call_tool('server_call_tool', telemetryData).catch(() => {});- await capture('server_prompt_retrieved', { + void capture('server_prompt_retrieved', { prompt_id: prompt.id, prompt_title: prompt.title, category: prompt.categories[0] || 'uncategorized', author: prompt.author, verified: prompt.verified - }); + }).catch(() => {});- capture('server_request_error', { + void capture('server_request_error', { error: errorMessage - }); + }).catch(() => {});Also applies to: 761-762, 772-773, 784-785, 805-812, 825-826, 927-927, 1030-1032
792-817: Avoid extra disk read; return prompt metadata from tool insteadThe handler lazily imports and reads prompts data again just to emit analytics. Push prompt metadata through the tool response (e.g., set result._meta = { id, title, categories[0], author, verified }) during the get_prompt path and use that here. This removes redundant I/O and keeps concerns separated.
src/tools/prompts.ts (4)
214-216: Don’t block on analytics bookkeepingPrompt usage tracking shouldn’t delay the user-visible response. Fire-and-forget and swallow errors.
- await usageTracker.markPromptUsed(promptId, prompt.categories[0] || 'uncategorized'); + void usageTracker.markPromptUsed(promptId, prompt.categories[0] || 'uncategorized').catch(() => {});
292-301: Sort within groups for stable, meaningful listingSurface verified prompts first, then by votes, then alphabetically. Improves UX consistency.
preferredOrder.forEach(tag => { if (groupedPrompts.has(tag)) { - const tagPrompts = groupedPrompts.get(tag)!; + const tagPrompts = groupedPrompts.get(tag)!.slice().sort((a, b) => { + if (a.verified !== b.verified) return a.verified ? -1 : 1; + if (b.votes !== a.votes) return b.votes - a.votes; + return a.title.localeCompare(b.title); + }); response += `${tag}:\n`; tagPrompts.forEach(prompt => { response += `${promptNumber}. ${prompt.title}\n`;
306-317: Avoid leaking internal IDs in rendered text; prefer _meta mappingThe HTML comment trick may still surface in some clients. Consider returning the number→ID map in ServerResult._meta (e.g., _meta: { promptMap }) and omit the comment entirely from the user-facing text.
74-119: Optional: validate params with the shared Zod schemaSince this tool can be called programmatically too, parse params with GetPromptsArgsSchema to guarantee consistent validation regardless of the caller. Keep existing friendly error messages for missing/invalid action.
If you want, I can wire in GetPromptsArgsSchema.safeParse and preserve your current error strings.
src/utils/usageTracker.ts (1)
365-372: Defensive read of onboarding statePersisted state can be partial or stale. Coalesce each field to defaults to avoid undefined math/logic later.
- const stored = await configManager.getValue('onboardingState'); - return stored || { - promptsUsed: false, - attemptsShown: 0, - lastShownAt: 0 - }; + const stored = await configManager.getValue('onboardingState'); + return { + promptsUsed: !!stored?.promptsUsed, + attemptsShown: Number.isFinite(stored?.attemptsShown) ? stored.attemptsShown : 0, + lastShownAt: Number.isFinite(stored?.lastShownAt) ? stored.lastShownAt : 0 + };
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
⛔ Files ignored due to path filters (1)
package-lock.jsonis excluded by!**/package-lock.json
📒 Files selected for processing (17)
.gitignore(1 hunks).serena/.gitignore(1 hunks).serena/memories/code_style_conventions.md(1 hunks).serena/memories/project_overview.md(1 hunks).serena/memories/suggested_commands.md(1 hunks).serena/memories/task_completion_checklist.md(1 hunks).serena/project.yml(1 hunks)FAQ.md(2 hunks)README.md(2 hunks)config_backup_before_onboarding_test.json(1 hunks)docs/index.html(1 hunks)src/data/onboarding-prompts.json(1 hunks)src/server.ts(5 hunks)src/tools/prompts.ts(1 hunks)src/utils/usageTracker.ts(2 hunks)test/test-binary-file-detection.js(0 hunks)test_onboarding.md(1 hunks)
💤 Files with no reviewable changes (1)
- test/test-binary-file-detection.js
✅ Files skipped from review due to trivial changes (5)
- .gitignore
- .serena/memories/project_overview.md
- .serena/memories/code_style_conventions.md
- .serena/.gitignore
- .serena/memories/task_completion_checklist.md
🧰 Additional context used
🧬 Code graph analysis (5)
.serena/memories/suggested_commands.md (1)
setup-claude-server.js (1)
setup(614-854)
src/utils/usageTracker.ts (1)
src/config-manager.ts (1)
configManager(219-219)
config_backup_before_onboarding_test.json (2)
src/config-manager.ts (1)
ConfigManager(29-216)test/test-negative-offset-readfile.js (1)
setup(28-48)
src/server.ts (4)
src/tools/schemas.ts (1)
GetPromptsArgsSchema(127-131)src/tools/prompts.ts (2)
getPrompts(74-129)loadPromptsData(50-69)src/utils/capture.ts (1)
capture(277-284)src/utils/usageTracker.ts (2)
usageTracker(488-488)shouldShowOnboarding(384-414)
src/tools/prompts.ts (2)
src/types.ts (1)
ServerResult(51-55)src/utils/usageTracker.ts (1)
usageTracker(488-488)
🪛 YAMLlint (1.37.1)
.serena/project.yml
[error] 25-25: trailing spaces
(trailing-spaces)
🪛 markdownlint-cli2 (0.17.2)
test_onboarding.md
36-36: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
🔇 Additional comments (6)
docs/index.html (1)
1615-1618: Nice onboarding nudge—LGTMClear CTA that ties into the new prompts flow. No issues.
FAQ.md (1)
25-25: TOC entry added—LGTMAnchor matches the section title below.
README.md (2)
25-25: TOC: Getting Started—LGTM
382-403: Getting Started matches the onboarding feature—LGTMGood examples and user phrasing align with the prompts catalog.
.serena/memories/suggested_commands.md (1)
1-56: Developer quick-reference—LGTMHandy and scoped to docs. No runtime impact.
src/data/onboarding-prompts.json (1)
1-123: Packaging and loader paths are correctly configured
The build script createsdist/dataand copiessrc/data/onboarding-prompts.jsonthere, andloadPromptsDatausespath.join(__dirname, '..', 'data', 'onboarding-prompts.json')—which points todist/data/onboarding-prompts.jsonat runtime. No further action required.
- Replace /Users/fiberta/ paths with ~/ placeholders - Add (replace with your path) instructions for clarity - Improve cross-platform compatibility in onboarding prompts - Addresses coderabbitai[bot] review feedback
- Uncomment the cache check in loadPromptsData() - Improves performance by avoiding redundant file reads - Cache is properly cleared via clearCache() function when needed
1 participant
Summary by CodeRabbit
New Features
Documentation
Chores
Tests