This skill reads all meta-prompt files from docs/prompts/_meta/ and executes each one to regenerate the corresponding workspace-tailored review prompt. Run this after setting up a new workspace or after making changes to the meta-prompts themselves.

Step 1: Identify All Meta-Prompts

List all markdown files in docs/prompts/_meta/. Verify the following meta-prompts are present:

Meta Prompt File	Target Output File
`optimize-code-review.md`	`docs/prompts/code-review-prompt.md`
`optimize-spring-cleaning.md`	`docs/prompts/spring-cleaning-prompt.md`
`optimize-readme.md`	`docs/prompts/readme-generation-prompt.md`
`optimize-architecture-review.md`	`docs/prompts/architecture-review.md`
`optimize-feature-complete.md`	`docs/prompts/feature-complete.md`
`optimize-feature-plan.md`	`docs/prompts/feature-plan-prompt.md`

If any meta-prompt is missing, warn the user and skip only that target.

Step 2: Execute Each Meta-Prompt

For each meta-prompt file found:

Read the content of the meta-prompt file.
Execute the instructions within the meta-prompt strictly:
- Each meta-prompt instructs you to analyze the current workspace (structure, tech stack, patterns, dependencies) and then rewrite a target prompt file to be codebase-aware.
- Identify the correct target output file as specified in the meta-prompt content.
Write the generated, workspace-tailored prompt to the specified target file, overwriting the existing scaffold/placeholder content.

[!NOTE] The meta-prompts are independent of each other. Parallelizing their execution is encouraged for efficiency when possible.

Step 3: Verify Output

After executing all meta-prompts:

Confirm each target output file listed in Step 1 has been regenerated with workspace-specific content (not just the generic scaffold placeholder).
Verify the new prompts are saved correctly by reading the first few lines of each target file.

Step 4: Confirm Completion

Report to the user:

✅ List all successfully regenerated prompt files
⚠️ List any meta-prompts that were skipped or failed, with the reason

Más de este repositorio

mismo repositorio

setup-reviews

ThMoJe/antigravity-self-evolving-reviews

Initializes the workspace with self-evolving reviews configurations, templates, prompts, and report directories.

2026-06-094

generate-readme

ThMoJe/antigravity-self-evolving-reviews

Audits the codebase and updates the root README.md to reflect the latest status, architectures, and guidelines.

2026-06-094

run-architecture-review

ThMoJe/antigravity-self-evolving-reviews

Perform a full-spectrum architectural audit of the system architecture, dependencies, and patterns.

2026-06-094

run-code-review

ThMoJe/antigravity-self-evolving-reviews

Perform a high-fidelity code review of the workspace or staged changes based on generated instructions.

2026-06-094

run-config-layer-audit

ThMoJe/antigravity-self-evolving-reviews

Audits and harmonizes all configuration layers (GEMINI.md, Skills, docs, MCP servers) to resolve inconsistencies.

2026-06-094

run-feature-complete

ThMoJe/antigravity-self-evolving-reviews

Updates documentation and guides, verifying the feature implementation before marking it complete.

2026-06-094

Generate Review Prompts Skill

Step 1: Identify All Meta-Prompts

List all markdown files in docs/prompts/_meta/. Verify the following meta-prompts are present:

Meta Prompt File

Target Output File

optimize-code-review.md

docs/prompts/code-review-prompt.md

optimize-spring-cleaning.md

docs/prompts/spring-cleaning-prompt.md

optimize-readme.md

docs/prompts/readme-generation-prompt.md

optimize-architecture-review.md

docs/prompts/architecture-review.md

optimize-feature-complete.md

docs/prompts/feature-complete.md

optimize-feature-plan.md

docs/prompts/feature-plan-prompt.md

If any meta-prompt is missing, warn the user and skip only that target.

Step 2: Execute Each Meta-Prompt

For each meta-prompt file found:

Read the content of the meta-prompt file.

Execute the instructions within the meta-prompt strictly:

Each meta-prompt instructs you to analyze the current workspace (structure, tech stack, patterns, dependencies) and then rewrite a target prompt file to be codebase-aware.
Identify the correct target output file as specified in the meta-prompt content.

Write the generated, workspace-tailored prompt to the specified target file, overwriting the existing scaffold/placeholder content.

[!NOTE] The meta-prompts are independent of each other. Parallelizing their execution is encouraged for efficiency when possible.

Step 3: Verify Output

After executing all meta-prompts:

Confirm each target output file listed in Step 1 has been regenerated with workspace-specific content (not just the generic scaffold placeholder).

Verify the new prompts are saved correctly by reading the first few lines of each target file.

Step 4: Confirm Completion

Report to the user:

✅ List all successfully regenerated prompt files

⚠️ List any meta-prompts that were skipped or failed, with the reason