Skip to content

Releases: Theorvane/planguard

PlanGuard BYO-AI Explanation v1.0.4

Choose a tag to compare

@sjungwon03 sjungwon03 released this 03 Aug 00:02
8fa62d6

Marketplace metadata fix

Shortens the root GitHub Action description to 75 characters, satisfying GitHub Marketplace's under-125-character requirement.

The Action behavior and security boundary are unchanged.

PlanGuard BYO-AI Explanation v1.0.3

Choose a tag to compare

@sjungwon03 sjungwon03 released this 02 Aug 14:31
fdfe8e7

Canonical repository migration

PlanGuard is now published from Theorvane/planguard. All README, Marketplace, GitHub App, and workflow references now use the canonical Action coordinates:

- uses: Theorvane/planguard@v1

Existing sjungwon03/planguard links redirect through GitHub, but new workflows should use the Theorvane coordinate.

PlanGuard BYO-AI Explanation v1.0.2

Choose a tag to compare

@sjungwon03 sjungwon03 released this 02 Aug 13:04
a0133c0

Documentation release

Adds Google Gemini first-party OpenAI compatibility configuration and Claude guidance through OpenRouter's OpenAI-compatible API.

Gemini and OpenRouter endpoints were accepted by PlanGuard's actual HTTPS/public-IPv4 endpoint validator. Anthropic first-party Messages API is not called directly by this Chat Completions Action.

PlanGuard BYO-AI Explanation v1.0.1

Choose a tag to compare

@sjungwon03 sjungwon03 released this 02 Aug 12:59
8082867

Documentation release

Adds verified OpenAI Chat Completions-compatible provider configuration examples for OpenAI, xAI (Grok), OpenRouter, Groq, Together AI, Mistral AI, and DeepSeek.

Each provider endpoint was accepted by PlanGuard's HTTPS/public-IPv4 endpoint validator. Model availability, pricing, and account access remain provider-controlled.

See the updated README for endpoint and model configuration.

PlanGuard BYO-AI Explanation v1.0.0

Choose a tag to compare

@sjungwon03 sjungwon03 released this 02 Aug 11:58
12eebe9

GitHub Actions / Marketplace release

PlanGuard BYO-AI Explanation is a composite Action for producing an optional OpenAI-compatible explanation from a pre-sanitized Terraform plan artifact.

Security boundary

  • Terraform executes only in the separately trusted preparation job.
  • This Action receives only the sensitive-value-redacted JSON artifact.
  • The AI API key stays in the explanation job GitHub Secret.
  • AI cannot calculate or override deterministic policy, risk, or pass/fail results.

Usage

Use sjungwon03/planguard@v1 after Marketplace publication, or pin this immutable release commit for strict supply-chain control.

See examples/workflows/ai-terraform-review.yml and docs/marketplace/README.md.