Skip to content

Latest commit

 

History

History
283 lines (220 loc) · 7.92 KB

configuration.md

File metadata and controls

283 lines (220 loc) · 7.92 KB

Configuration Reference

DevsContext uses .devscontext.yaml for all configuration. This file should be placed in your project root.

File Location

DevsContext searches for .devscontext.yaml in the current directory and parent directories, similar to how git finds .gitconfig.

Environment Variables

Use ${VAR_NAME} or $VAR_NAME syntax to reference environment variables. This is required for sensitive values like API keys:

sources:
  jira:
    api_token: "${JIRA_API_TOKEN}"

Configuration Schema

sources

Configuration for data source adapters.

sources.jira

Field Type Default Description
enabled bool false Enable Jira adapter
base_url string "" Jira instance URL (e.g., https://company.atlassian.net)
email string "" Authentication email
api_token string "" Jira API token (use ${JIRA_API_TOKEN})
project string "" Default project key for queries
primary bool true Fetch first; share context with secondary sources

sources.fireflies

Field Type Default Description
enabled bool false Enable Fireflies adapter
api_key string "" Fireflies API key (use ${FIREFLIES_API_KEY})
primary bool false Whether this is a primary source

sources.docs

Field Type Default Description
enabled bool true Enable local docs adapter
paths list[string] ["./docs/"] Paths to documentation directories
standards_path string null Optional path to coding standards
architecture_path string null Optional path to architecture docs
primary bool false Whether this is a primary source
rag object null Optional RAG configuration (see below)
sources.docs.rag

Optional embedding-based search configuration. Requires pip install devscontext[rag].

Field Type Default Description
enabled bool false Enable RAG for semantic doc matching
embedding_provider string "local" Provider: local, openai, or ollama
embedding_model string "all-MiniLM-L6-v2" Model for generating embeddings
index_path string ".devscontext/doc_index.json" Path to embedding index file
top_k int 10 Number of similar sections to retrieve
similarity_threshold float 0.3 Minimum similarity score (0-1)

sources.slack

Field Type Default Description
enabled bool false Enable Slack adapter
bot_token string "" Slack bot token (use ${SLACK_BOT_TOKEN})
channels list[string] [] Channel names to search
include_threads bool true Fetch full threads for matches
max_messages int 20 Max messages per search (1-100)
lookback_days int 30 Days to search back (1-90)
primary bool false Whether this is a primary source

sources.gmail

Field Type Default Description
enabled bool false Enable Gmail adapter
credentials_path string "" Path to OAuth2 credentials JSON
token_path string ".devscontext/gmail_token.json" Path for OAuth2 refresh token
search_scope string "newer_than:30d" Gmail search scope filter
max_results int 10 Max emails to return (1-50)
labels list[string] ["INBOX"] Labels to search within
primary bool false Whether this is a primary source

synthesis

Configuration for context synthesis.

Field Type Default Description
plugin string "llm" Synthesis plugin: llm, template, passthrough
provider string "anthropic" LLM provider: anthropic, openai, ollama
model string "claude-haiku-4-5" Model name/ID
api_key string null API key (use env var, e.g., ${ANTHROPIC_API_KEY})
max_output_tokens int 3000 Max tokens in synthesized output (100-10000)
temperature float 0.0 LLM temperature (0.0-2.0)
prompt_template string null Path to custom prompt template
template_path string null Jinja2 template path (for template plugin)

cache

In-memory cache configuration.

Field Type Default Description
enabled bool true Enable caching
ttl_minutes int 15 Cache TTL in minutes (1-1440)
max_size int 100 Maximum cache entries

agents

Background agent configuration.

agents.preprocessor

Field Type Default Description
enabled bool false Enable preprocessing agent
jira_status string "Ready for Development" Jira status that triggers processing
jira_project string or list "" Project key(s) to watch
context_ttl_hours int 24 How long pre-built context is valid (1-168)

agents.preprocessor.trigger

Field Type Default Description
type string "polling" Trigger type (only polling supported)
poll_interval_minutes int 5 Poll interval (1-60 minutes)

storage

Persistent storage configuration.

Field Type Default Description
path string ".devscontext/cache.db" SQLite database path

Example Configurations

Minimal (Jira + Local Docs)

sources:
  jira:
    enabled: true
    base_url: "https://company.atlassian.net"
    email: "${JIRA_EMAIL}"
    api_token: "${JIRA_API_TOKEN}"

  docs:
    enabled: true
    paths:
      - "./docs"
      - "./CLAUDE.md"

synthesis:
  provider: "anthropic"
  model: "claude-haiku-4-5"

Full Configuration

sources:
  jira:
    enabled: true
    base_url: "https://company.atlassian.net"
    email: "${JIRA_EMAIL}"
    api_token: "${JIRA_API_TOKEN}"
    project: "PROJ"
    primary: true

  fireflies:
    enabled: true
    api_key: "${FIREFLIES_API_KEY}"
    primary: false

  docs:
    enabled: true
    paths:
      - "./docs"
      - "./CLAUDE.md"
    standards_path: "./docs/standards"
    architecture_path: "./docs/architecture"
    primary: false
    rag:
      enabled: true
      embedding_provider: "local"
      embedding_model: "all-MiniLM-L6-v2"
      index_path: ".devscontext/doc_index.json"
      top_k: 10
      similarity_threshold: 0.3

  slack:
    enabled: true
    bot_token: "${SLACK_BOT_TOKEN}"
    channels:
      - "engineering"
      - "payments-team"
    include_threads: true
    max_messages: 20
    lookback_days: 30

  gmail:
    enabled: false
    credentials_path: "${GMAIL_CREDENTIALS_PATH}"
    token_path: ".devscontext/gmail_token.json"
    search_scope: "newer_than:30d"
    max_results: 10
    labels:
      - "INBOX"

synthesis:
  plugin: "llm"
  provider: "anthropic"
  model: "claude-haiku-4-5"
  max_output_tokens: 3000
  temperature: 0.0

cache:
  enabled: true
  ttl_minutes: 15
  max_size: 100

agents:
  preprocessor:
    enabled: true
    jira_status: "Ready for Development"
    jira_project: "PROJ"
    context_ttl_hours: 24
    trigger:
      type: "polling"
      poll_interval_minutes: 5

storage:
  path: ".devscontext/cache.db"

With RAG Enabled

sources:
  docs:
    enabled: true
    paths:
      - "./docs"
    rag:
      enabled: true
      embedding_provider: "local"  # Uses sentence-transformers
      embedding_model: "all-MiniLM-L6-v2"
      top_k: 10
      similarity_threshold: 0.3

synthesis:
  provider: "anthropic"
  model: "claude-haiku-4-5"

After configuring RAG, build the index:

pip install devscontext[rag]
devscontext index-docs