[COORD-004] Implement agent profiles #144

Closed
opened 2026-01-31 21:03:22 +00:00 by jason.woltje · 0 comments
Owner

Objective

Implement agent profile system defining capabilities, costs, and context limits for each model.

Implementation Details

Create the agent capability matrix from Part 1:

Agent profiles define:

  • context_limit: Maximum tokens for the agent
  • cost_per_mtok: Cost per million tokens
  • capabilities: List of difficulty levels the agent can handle
  • best_for: Description of optimal use cases

Profiles needed:

  • opus: 200K limit, high/medium/low capabilities, high cost
  • sonnet: 200K limit, medium/low capabilities, medium cost
  • haiku: 200K limit, low capability, low cost
  • glm: 128K limit, medium/low capabilities, free (self-hosted)
  • minimax: 128K limit, low capability, free (self-hosted)

Context Estimate

  • Files to modify: 2 (agent_assignment.py, models.py)
  • Implementation complexity: low (10,000 tokens)
  • Test requirements: low (5,000 tokens)
  • Documentation: light (2,000 tokens)
  • Total estimated: 31,200 tokens
  • Recommended agent: haiku

Difficulty

low

Dependencies

  • Blocked by: #145 (COORD-003)
  • Blocks: #147 (COORD-005)

Acceptance Criteria

  • Agent profile data structure defined
  • All 5 agent profiles implemented (opus, sonnet, haiku, glm, minimax)
  • Profile fields include: context_limit, cost_per_mtok, capabilities, best_for
  • Profile validation logic ensures consistency
  • Unit tests for profile data (85% coverage minimum)

Testing Requirements

  • Test profile data completeness
  • Test profile validation logic
  • Test capability lookups
  • Coverage: 85% minimum
## Objective Implement agent profile system defining capabilities, costs, and context limits for each model. ## Implementation Details Create the agent capability matrix from Part 1: Agent profiles define: - context_limit: Maximum tokens for the agent - cost_per_mtok: Cost per million tokens - capabilities: List of difficulty levels the agent can handle - best_for: Description of optimal use cases Profiles needed: - opus: 200K limit, high/medium/low capabilities, high cost - sonnet: 200K limit, medium/low capabilities, medium cost - haiku: 200K limit, low capability, low cost - glm: 128K limit, medium/low capabilities, free (self-hosted) - minimax: 128K limit, low capability, free (self-hosted) ## Context Estimate - Files to modify: 2 (agent_assignment.py, models.py) - Implementation complexity: low (10,000 tokens) - Test requirements: low (5,000 tokens) - Documentation: light (2,000 tokens) - **Total estimated: 31,200 tokens** - **Recommended agent: haiku** ## Difficulty low ## Dependencies - Blocked by: #145 (COORD-003) - Blocks: #147 (COORD-005) ## Acceptance Criteria - [ ] Agent profile data structure defined - [ ] All 5 agent profiles implemented (opus, sonnet, haiku, glm, minimax) - [ ] Profile fields include: context_limit, cost_per_mtok, capabilities, best_for - [ ] Profile validation logic ensures consistency - [ ] Unit tests for profile data (85% coverage minimum) ## Testing Requirements - Test profile data completeness - Test profile validation logic - Test capability lookups - Coverage: 85% minimum
jason.woltje added the phase-2apiapip0 labels 2026-01-31 21:03:22 +00:00
jason.woltje added this to the M4.1-Coordinator (0.0.4) milestone 2026-01-31 21:10:03 +00:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: mosaic/stack#144