ros/claude-task-master
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(ai): Enhance Perplexity research calls & fix docs examples

Browse Source 
Improves the quality and relevance of research-backed AI operations:
- Tweaks Perplexity AI calls to use max input tokens (8700), temperature 0.1, high context size, and day-fresh search recency.
- Adds a system prompt to guide Perplexity research output.

Docs:
- Updates CLI examples in taskmaster.mdc to use ANSI-C quoting ($'...') for multi-line prompts, ensuring they work correctly in bash/zsh.
This commit is contained in:
Eyal Toledano
2025-04-14 17:09:06 -04:00
parent a0663914e6
commit 2c2e60ad55
8 changed files with 211 additions and 37 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
scripts/modules/ai-services.js
View File

@@ -709,12 +709,24 @@ Include concrete code examples and technical considerations where relevant.`;
const researchResponse = await perplexityClient.chat.completions.create({
model: PERPLEXITY_MODEL,
messages: [
{
role: 'system',
content: `You are a helpful assistant that provides research on current best practices and implementation approaches for software development.
You are given a task and a description of the task.
You need to provide a list of best practices, libraries, design patterns, and implementation approaches that are relevant to the task.
You should provide concrete code examples and technical considerations where relevant.`
},
{
role: 'user',
content: researchQuery
}
],
temperature: 0.1 // Lower temperature for more factual responses
temperature: 0.1, // Lower temperature for more factual responses
max_tokens: 8700, // Respect maximum input tokens for Perplexity (8719 max)
web_search_options: {
search_context_size: 'high'
},
search_recency_filter: 'day' // Filter for results that are as recent as today to capture new releases
});
const researchResult = researchResponse.choices[0].message.content;
@@ -814,7 +826,7 @@ Note on dependencies: Subtasks can depend on other subtasks with lower IDs. Use
anthropic,
{
model: session?.env?.ANTHROPIC_MODEL || CONFIG.model,
max_tokens: session?.env?.MAX_TOKENS || CONFIG.maxTokens,
max_tokens: 8700,
temperature: session?.env?.TEMPERATURE || CONFIG.temperature,
system: systemPrompt,
messages: [{ role: 'user', content: userPrompt }]
@@ -1328,7 +1340,12 @@ Include concrete code examples and technical considerations where relevant.`;
content: researchQuery
}
],
temperature: 0.1 // Lower temperature for more factual responses
temperature: 0.1, // Lower temperature for more factual responses
max_tokens: 8700, // Respect maximum input tokens for Perplexity (8719 max)
web_search_options: {
search_context_size: 'high'
},
search_recency_filter: 'day' // Filter for results that are as recent as today to capture new releases
});
const researchResult = researchResponse.choices[0].message.content;