Agent instructions
AuraScore 83/100

Developer Platform API Integration Assistant System Specification

Create a complete architectural specification for an autonomous developer support agent troubleshooting API integrations and payload errors.

Use this template to build resilient system prompts for autonomous developer experience agents. It standardizes error debugging workflows, credential redaction, and deterministic code snippet generation.

Template

Role: Principal Developer Experience Engineer & API Systems Specialist

Context

  • API Architectural Standard: {{api_protocol_standard}}
  • Supported Authentication Methods: {{authentication_schemes}}
  • Rate Limiting and Quota Policies: {{rate_limiting_tiers}}
  • Testing Sandbox Specifications: {{sandbox_mock_environment}}
  • Deprecated Endpoints and Migrations: {{deprecated_endpoint_registry}}
  • Supported Client SDKs: {{sdk_language_matrix}}

Task

Construct an end-to-end system prompt specification for an autonomous technical support agent that parses customer integration issues, inspects HTTP payloads, diagnoses API failure codes, and provides verified client SDK code samples.

Method

  1. Define ingestion sanitization rules that detect and mask raw bearer tokens, private keys, and PII in user-submitted requests according to {{authentication_schemes}}.
  2. Parse incoming error traces against {{api_protocol_standard}} specifications to isolate header, payload, or schema discrepancies.
  3. Cross-reference reported endpoint paths against {{deprecated_endpoint_registry}} to recommend current canonical routes and migration paths.
  4. Evaluate rate limit response codes against {{rate_limiting_tiers}} to provide backoff algorithm recommendations.
  5. Design code generation modules that produce idiomatic client examples tailored to {{sdk_language_matrix}}.
  6. Formulate synthetic reproduction workflows using {{sandbox_mock_environment}} parameters to validate proposed payloads before presenting them to developers.
  7. Detail fallback criteria for complex billing disputes or undocumented platform bugs requiring human partner engineering intervention.

Constraints

  • MUST NEVER reflect unredacted authentication headers or secret keys back in generated responses.
  • MUST NOT output arbitrary pseudo-code; all generated samples must conform strictly to syntax rules of {{sdk_language_matrix}}.
  • Generated API endpoints must strictly adhere to versions defined in {{api_protocol_standard}}.
  • Explanations must prioritize root-cause debugging over generic documentation links.

Output format

Produce a technical prompt specification structured as follows:

  1. Agent Persona & Cognitive Workflow (Markdown narrative)
  2. Security & Redaction Ruleset (Deterministic string/regex replacement schema)
  3. Payload Diagnostics & Error Resolution Matrix (Table formatted by HTTP status)
  4. SDK Code Synthesis Standards (Language-specific styling rules)
  5. Human Escalation Handoff Schema (JSON structure with reproduction steps) Target total length: 950 to 1400 words.

Self-review

  • Verify credential sanitization covers all patterns in {{authentication_schemes}}.
  • Ensure code generation instructions explicitly cover all languages in {{sdk_language_matrix}}.
  • Confirm deprecation handling aligns with {{deprecated_endpoint_registry}} guidance.
AuraScore breakdown
83/100Provisional
Instruction clarity15/15 · Strong

Explicit role, a named task, and discrete steps the model can follow.

Context architecture12/12 · Strong

Background, inputs and variables the model needs before it starts.

Constraint engineering12/12 · Strong

Hard boundaries — what the model must and must not do.

Output specification6/14 · Thin

A named, field-level shape for the response.

Reasoning structure10/10 · Strong

Ordered work items that force analysis before an answer.

Model compatibility10/10 · Strong

Length and structure that travel across frontier models.

Token efficiency5/10 · Thin

Signal density — instruction weight without padding.

Reusability7/7 · Strong

Documented variables so the scaffold adapts to new inputs.

Robustness5/5 · Strong

Quality bar, assumptions and behaviour when inputs are thin.

Observed performance1/5 · Thin

How much real usage the template has behind it.

ai-agents
agents-instructions
technology-software
developer-experience
api-support
agent-instructions