Use when adding new error messages to React, or seeing "unknown error code" warnings.
npx skills add manarouei/agent-skills --skill "code-implement"
Install specific skill from multi-skill repository
# Description
Git diff of changes
# SKILL.md
name: code-implement
version: "1.0.0"
description: Implement node from API documentation using LLM. For TYPE2 sources only. Generates implementation based on trace-mapped schema.
Contract
autonomy_level: IMPLEMENT
side_effects: [fs, net]
timeout_seconds: 600
retry:
policy: none
max_retries: 0
idempotency:
required: true
key_spec: "correlation_id"
max_fix_iterations: 3
Sync Celery Constraints (MANDATORY)
sync_celery:
requires_sync_execution: true
forbids_async_dependencies: true
requires_timeouts_on_external_calls: true
forbids_background_tasks: true
input_schema:
type: object
required: [correlation_id, source_type, node_schema, trace_map, allowlist]
properties:
correlation_id:
type: string
source_type:
type: string
const: TYPE2
node_schema:
type: object
trace_map:
type: object
allowlist:
type: object
output_schema:
type: object
required: [files_modified, implementation_notes, assumptions_used]
properties:
files_modified:
type: array
items: { type: string }
implementation_notes:
type: array
assumptions_used:
type: array
description: List of trace_map ASSUMPTION entries used
required_artifacts:
- name: implementation_log.json
type: json
description: Implementation decisions
- name: files.diff
type: diff
description: Git diff of changes
failure_modes: [parse_error, scope_violation, validation_error, trace_incomplete]
depends_on: [node-scaffold]
Code Implement
Generate implementation from API documentation (TYPE2 only).
SCOPE ENFORCEMENT
ALL file modifications MUST be within allowlist patterns.
TRACE MAP REQUIREMENT
Every implemented operation MUST reference trace_map entries.
If an operation requires an ASSUMPTION from trace_map:
1. Document in assumptions_used output
2. Add TODO comment in code
3. Flag for human verification
Implementation Process
For each operation in schema:
- Read trace_map for operation parameters
- Construct API request from documented patterns
- Add authentication from credentials
- Handle response parsing
- Add error handling
BaseNode Method Template
def execute(self) -> List[List[NodeExecutionData]]:
input_data = self.get_input_data()
results = []
for item in input_data:
operation = self.get_node_parameter("operation", 0)
if operation == "create":
result = self._execute_create(item)
elif operation == "get":
result = self._execute_get(item)
# ... more operations
results.append(result)
return [results]
def _execute_create(self, item: NodeExecutionData) -> NodeExecutionData:
credentials = self.get_credentials("apiCredential")
# Implementation based on trace_map evidence
...
DO NOT
- Invent parameters without trace_map evidence
- Skip ASSUMPTION verification
- Implement undocumented endpoints
Artifacts Emitted
artifacts/{correlation_id}/implementation_log.jsonartifacts/{correlation_id}/files.diff
# Supported AI Coding Agents
This skill is compatible with the SKILL.md standard and works with all major AI coding agents:
Learn more about the SKILL.md standard and how to use these skills with your preferred AI coding agent.