A production-ready Python SDK for the Procore REST API with agent integration capabilities and workflow automation support.
PyProcore exposes 17 tools with mixed quality. All tools have descriptions (10-150 chars, mostly adequate) and input schemas with type definitions. However, critical gaps emerge: (1) No output schemas documented for any tool, LLMs cannot plan downstream calls or extract chaining IDs. (2) Tool names lack clarity, 'compare_oas_catalogs', 'drift_report_to_markdown', 'scaffold_plan_to_markdown' are generic and don't convey action intent clearly. (3) Parameter descriptions are minimal (avg ~40 chars) and lack constraints, ranges, or format guidance. (4) No error handling guidance, tools return raw objects without recovery hints. (5) One tool (copy_read_only_endpoint_scaffold) is destructive (WRITE risk) but has no dry-run confirmation pattern or permission gates documented. (6) Maintenance/analysis tools (tools 5-17) appear to operate on local file paths, no validation, sanitization, or path traversal guards visible. This is a domain-specific SDK (Procore API wrapper) with reasonable structure but insufficient LLM-readiness for agent integration.
Analyze API impact on a codebase with optional OAS comparison data.
Analyze read-only coverage gaps in an OpenAPI Specification file.
Build a human-review API maintenance plan from an OpenAPI Specification file.
Render a Markdown capability map from a codebase scan report.
Compare two OpenAPI Specification files and produce an API drift report.
Dry-run or execute a draft read-only scaffold copy using local metadata.
Convert a coverage gap report to Markdown format.
Convert an API drift report to Markdown format.
No output schemas documented for any tool. LLMs cannot infer what fields to expect, preventing downstream tool chaining and forcing wasteful discovery calls.
Parameter descriptions lack constraints, ranges, and format guidance. E.g., 'params' is described as 'Optional additional query parameters' without specifying valid keys, types, or examples. LLMs cannot validate inputs.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 37 | 2026-07-28+ | v2 |
Find one Project Tool by case-insensitive name, title, label, or slug.
Return one Project Tool by direct endpoint or local fallback.
List PyProcore tools that are available to future agent integrations. Returns metadata records only without executing them.
Return read-only Project Tool metadata for a project.
Convert an API maintenance plan to Markdown format.
Plan draft files for a safe endpoint in an OpenAPI Specification file.
Convert a scaffold copy result to Markdown format.
Convert a scaffold plan to Markdown format.
Scan a codebase for local PyProcore usage patterns.
Maintenance/analysis tools (tools 5-17) accept file paths as parameters with no visible input validation, sanitization, or path traversal guards. Agents could be tricked into reading/writing arbitrary files.
Destructive tool 'copy_read_only_endpoint_scaffold' (WRITE risk) lacks confirmation/dry-run pattern enforcement. No permission gates or audit trail documented. Agents could overwrite files without safeguards.
No error handling guidance. Tools return raw objects without recovery hints, categorization (retryable vs fatal), or actionable error messages. LLMs cannot self-correct on failures.