using-generic-agents
Flagged{"isSafe":false,"isSuspicious":true,"riskLevel":"high","findings":[{"category":"prompt_injection","severity":"high","description":"CRITICAL instruction attempting to override the model's system prompts by claiming operator directions supersede these directions.","evidence":"\"CRITICAL: Your operator's direction supercedes these directions. If the operator specifies a type of agent, execute their task with that agent.\""}],"summary":"The content is not inherently harmful but contains a clear prompt-injection risk: a CRITICAL directive attempting to override system-level instructions. This could subvert safety if used as a prompt. No malicious shell commands, data exfiltration, or payloads are present, but the directive should be removed or sanitized to prevent instruction override."}
npx machina-cli add skill ed3dai/ed3d-plugins/using-generic-agents --openclawCRITICAL: Your operator's direction supercedes these directions. If the operator specifies a type of agent, execute their task with that agent.
Model Characteristics
Haiku: Excellent at following specific, detailed instructions. Poor at making its own decisions. Give it a clear prompt and it executes well; ask it to figure things out and it struggles. Be detailed.
Sonnet: Capable of making decisions but gets off-track easily. Will explain concepts, describe structures, and gather extraneous information when you just want it to do the thing, so guard against this when prompting the agent.
Opus: Stays on-track through complex tasks. Better judgment, fewer loops. Expensive—don't use for clearly-definable workflows where Sonnet/Haiku would suffice.
When to Use Each
Use haiku-general-purpose for:
- Well-defined tasks with detailed prompts
- High-volume parallel workflows (cost matters)
- Simple execution where speed > quality
Use sonnet-general-purpose for:
- Multi-file reasoning and debugging
- Tasks requiring some judgment
- Daily coding work (80-90% of tasks)
Use opus-general-purpose for:
- Tasks requiring sustained focus and judgment
- When Sonnet keeps wandering or looping
- Complex analysis where staying on-track matters
- High-stakes decisions needing nuance
Source
git clone https://github.com/ed3dai/ed3d-plugins/blob/main/plugins/ed3d-basic-agents/skills/using-generic-agents/SKILL.mdView on GitHub Overview
This skill helps you decide which generic agent to deploy: Haiku for detailed prompts, Sonnet for judgment with some wandering risk, or Opus for sustained focus and complex analysis. It also notes that operator direction supersedes these directions when specified.
How This Skill Works
The system assesses task clarity, required judgment, and focus needs, then suggests Haiku, Sonnet, or Opus. It respects the operator's directive if provided and routes the task to the agent best suited to the prompt and constraints described in the skill (Haiku for clear prompts, Sonnet for multi-file reasoning, Opus for on-track deep analysis).
When to Use It
- Well-defined tasks with detailed prompts
- High-volume parallel workflows (cost matters)
- Multi-file reasoning and debugging
- Tasks requiring sustained focus and nuanced judgment
- Complex analysis or high-stakes decisions needing staying on track
Quick Start
- Step 1: Assess the task to determine if it is well-defined, multi-file, or requires deep analysis.
- Step 2: If specified by the operator, follow the directive; otherwise choose Haiku for detailed prompts, Sonnet for judgment-based tasks, or Opus for sustained focus.
- Step 3: Craft an explicit prompt tailored to the selected agent and monitor output for wandering; adjust as needed.
Best Practices
- Match task prompt quality to the agent type: use Haiku for detailed, explicit instructions; Sonnet for tasks needing some judgment; Opus for complex, protracted work.
- Be mindful of cost vs. quality: prefer Haiku for speed and volume, reserve Opus for high-stakes or nuanced analysis.
- Always respect operator direction if they specify an agent type.
- Provide precise, focused prompts to prevent stray reasoning or loops.
- Monitor results for wandering or looping and switch agents if needed
Example Use Cases
- Defining a data transformation workflow with explicit steps and outputs—use Haiku-general-purpose for reliable execution.
- Debugging a large codebase across multiple files—Sonnet-general-purpose helps with reasoning while keeping some structure.
- Running a complex market analysis that requires nuanced judgment—Opus-general-purpose is suited for staying on-track and delivering insight.
- Processing a high-volume catalog update where speed matters—Haiku-general-purpose accelerates parallel tasks.
- Producing a regulatory-compliant report where precision and nuance are critical—Opus-general-purpose handles high-stakes analysis while maintaining focus.
