Library — DistroSource

Everything digital. One source.

9 resources

Prompt · Advanced · 12 min
Agent evaluation rubric — variations
Produce several distinct agent evaluation rubric options so the best can be chosen or tested, each built on a different angle rather than reworded copies.
- Fill in: {{agent}}, {{risks}}, {{volume}}, {{pass}}
Variables 4ClaudeChatGPT
Prompt · Advanced · 8 min
Agent evaluation rubric — review and improve
Review an existing agent evaluation rubric against the standard a strong builder who designs agent system prompts that hold up in production would hold it to, then return an improved version.
- Fill in: {{agent}}, {{risks}}, {{volume}}, {{pass}}, {{draft}}
Variables 5ClaudeChatGPT
Prompt · Advanced · 8 min
Agent evaluation rubric
Write the rubric and test set used to evaluate an agent's system prompt before it goes live.
- Write the rubric and test set used to evaluate an agent's system prompt before it goes live. Written for AI agents work: the context, steps and output format are fixed so results come back in the same shape every time.
- Fill in: {{agent}}, {{risks}}, {{volume}}, {{pass}}
Variables 4ClaudeChatGPT
Prompt · Advanced · 12 min
Content moderation agent — variations
Produce several distinct content moderation agent options so the best can be chosen or tested, each built on a different angle rather than reworded copies.
- Fill in: {{policy}}, {{content}}, {{actions}}, {{bias}}
Variables 4ClaudeChatGPT
Prompt · Advanced · 8 min
Content moderation agent — review and improve
Review an existing content moderation agent against the standard a strong builder who designs agent system prompts that hold up in production would hold it to, then return an improved version.
- Fill in: {{policy}}, {{content}}, {{actions}}, {{bias}}, {{draft}}
Variables 5ClaudeChatGPT
Prompt · Advanced · 8 min
Content moderation agent
Define a system prompt for an agent that reviews user-submitted content against a policy and explains each decision.
- Define a system prompt for an agent that reviews user-submitted content against a policy and explains each decision. Written for AI agents work: the context, steps and output format are fixed so results come back in the same shape every time.
- Fill in: {{policy}}, {{content}}, {{actions}}, {{bias}}
Variables 4ClaudeChatGPT
Prompt · Advanced · 12 min
Code review agent — variations
Produce several distinct code review agent options so the best can be chosen or tested, each built on a different angle rather than reworded copies.
- Fill in: {{standard}}, {{focus}}, {{actions}}, {{cannot}}
Variables 4ClaudeChatGPT
Prompt · Advanced · 8 min
Code review agent — review and improve
Review an existing code review agent against the standard a strong builder who designs agent system prompts that hold up in production would hold it to, then return an improved version.
- Fill in: {{standard}}, {{focus}}, {{actions}}, {{cannot}}, {{draft}}
Variables 5ClaudeChatGPT
Prompt · Advanced · 8 min
Code review agent
Define a system prompt for an agent that reviews pull requests against a written standard and never approves what it cannot verify.
- Define a system prompt for an agent that reviews pull requests against a written standard and never approves what it cannot verify. Written for AI agents work: the context, steps and output format are fixed so results come back in the same shape every time.
- Fill in: {{standard}}, {{focus}}, {{actions}}, {{cannot}}
Variables 4ClaudeChatGPT

Search