@promptscript/cli 1.4.0 → 1.4.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -2,7 +2,7 @@
2
2
 
3
3
  **One source of truth for all your AI coding assistants.**
4
4
 
5
- Write AI instructions once in PromptScript, compile to GitHub Copilot, Claude Code, Cursor, Antigravity, and more.
5
+ Write AI instructions once in PromptScript, compile to 37 AI coding agents: GitHub Copilot, Claude Code, Cursor, and more.
6
6
 
7
7
  [![npm version](https://img.shields.io/npm/v/@promptscript/cli.svg)](https://www.npmjs.com/package/@promptscript/cli)
8
8
  [![CI](https://github.com/mrwogu/promptscript/actions/workflows/ci.yml/badge.svg)](https://github.com/mrwogu/promptscript/actions/workflows/ci.yml)
@@ -12,34 +12,19 @@ Write AI instructions once in PromptScript, compile to GitHub Copilot, Claude Co
12
12
 
13
13
  ## The Problem
14
14
 
15
- Every AI coding tool uses a different config format. As your organization grows, so does the chaos:
15
+ - 50 repos × 37 AI tools = **1,900 files** to keep in sync
16
+ - ❌ Security policy update? Manual changes across every repo
17
+ - ❌ Switching AI tools? Rewrite everything
18
+ - ❌ No audit trail, no inheritance, no validation
16
19
 
17
- ```
18
- 50 repos x 37 AI agents = 1850 files to maintain
19
-
20
- repo-1/CLAUDE.md repo-1/.cursorrules repo-1/.github/copilot-instructions.md
21
- repo-2/CLAUDE.md repo-2/.cursorrules repo-2/.github/copilot-instructions.md
22
- ...
23
- repo-50/CLAUDE.md repo-50/.cursorrules repo-50/.github/copilot-instructions.md
24
- ```
20
+ ## The Fix
25
21
 
26
- - A security policy update means editing hundreds of files by hand
27
- - Standards drift across teams and repos
28
- - Switching AI tools means rewriting everything
29
- - No audit trail, no validation, no reuse
22
+ PromptScript lets you **build prompts like code** - with inheritance, composition, parameterization, and compilation to any target format.
30
23
 
31
- ## The Solution
32
-
33
- PromptScript lets you **build prompts like code** — with inheritance, composition, parameterization, and compilation to any target format:
34
-
35
- ```
36
- .promptscript/project.prs --> prs compile --> CLAUDE.md
37
- --> .github/copilot-instructions.md
38
- --> .cursor/rules/project.mdc
39
- --> .agent/rules/project.md
40
- ```
41
-
42
- Update once, propagate everywhere. Version-controlled, validated, vendor-independent.
24
+ - Write once in `.prs`, compile to **all 37 agents**
25
+ - ✅ Update the source, propagates everywhere automatically
26
+ - Hierarchical inheritance like code, not copy-paste
27
+ - ✅ Full validation, audit trail, and version pinning
43
28
 
44
29
  ---
45
30
 
@@ -93,6 +78,14 @@ prs init --migrate
93
78
  "/review": "Security-focused code review"
94
79
  "/test": "Write unit tests with Vitest"
95
80
  }
81
+
82
+ @skills {
83
+ deploy: {
84
+ description: "Deploy service to production"
85
+ userInvocable: true
86
+ allowedTools: ["Bash", "Read"]
87
+ }
88
+ }
96
89
  ```
97
90
 
98
91
  Run `prs compile` and get correctly formatted output for every AI tool your team uses.
@@ -105,7 +98,7 @@ Run `prs compile` and get correctly formatted output for every AI tool your team
105
98
  | :---------------------- | :------------------------------------------- |
106
99
  | `prs init` | Initialize project with auto-detection |
107
100
  | `prs compile` | Compile to target AI tool formats |
108
- | `prs compile -w` | Watch mode recompile on changes |
101
+ | `prs compile -w` | Watch mode - recompile on changes |
109
102
  | `prs compile --dry-run` | Preview changes without writing files |
110
103
  | `prs validate` | Validate `.prs` files with detailed errors |
111
104
  | `prs diff` | Show diff between source and compiled output |
@@ -115,27 +108,31 @@ Run `prs compile` and get correctly formatted output for every AI tool your team
115
108
 
116
109
  ## Key Features
117
110
 
118
- - **Inheritance** build org-wide, team-level, and project-level configs that cascade like CSS
119
- - **Composition** reuse fragments across projects with `@use`
120
- - **Parameterized templates** `@inherit @stacks/node(port: 8080, db: "postgres")`
121
- - **Multi-target compilation** one source, any number of AI tools
122
- - **Watch mode** instant recompilation on file changes
123
- - **Overwrite protection** never accidentally clobbers hand-written files
124
- - **Validation** catch errors before they reach your AI tools
125
- - **Registry support** share configs via Git registries (private or public)
126
- - **AI-assisted migration** convert existing `CLAUDE.md`, `.cursorrules`, etc.
111
+ - **Inheritance** - build org-wide, team-level, and project-level configs that cascade like CSS
112
+ - **Composition** - reuse fragments across projects with `@use`
113
+ - **Parameterized templates** - `@inherit @stacks/node(port: 8080, db: "postgres")`
114
+ - **Skills** - define reusable AI skills with `SKILL.md` files, resource bundles, and tool permissions
115
+ - **Multi-target compilation** - one source, any number of AI tools
116
+ - **Watch mode** - instant recompilation on file changes
117
+ - **Overwrite protection** - never accidentally clobbers hand-written files
118
+ - **Validation** - catch errors before they reach your AI tools
119
+ - **Registry support** - share configs via Git registries (private or public)
120
+ - **AI-assisted migration** - convert existing `CLAUDE.md`, `.cursorrules`, etc.
121
+ - **Bundled language skill** - AI agents learn PromptScript syntax via injected SKILL.md
127
122
 
128
123
  ## Supported Targets
129
124
 
130
125
  | AI Tool | Output |
131
126
  | :--------------------- | :---------------------------------------------------- |
132
127
  | **GitHub Copilot** | `.github/copilot-instructions.md`, agents, prompts |
133
- | **Claude Code** | `CLAUDE.md`, skills, local memory |
128
+ | **Claude Code** | `CLAUDE.md`, `.claude/skills/*.md` |
134
129
  | **Cursor** | `.cursor/rules/*.mdc` |
135
130
  | **Google Antigravity** | `.agent/rules/*.md` |
136
131
  | **Factory AI** | `AGENTS.md`, `.factory/skills/`, `.factory/commands/` |
137
- | **OpenCode** | `agents.yaml` |
138
- | **Gemini CLI** | `GEMINI.md` |
132
+ | **OpenCode** | `OPENCODE.md`, `.opencode/commands/*.md` |
133
+ | **Gemini CLI** | `GEMINI.md`, `.gemini/commands/*.toml` |
134
+
135
+ Plus **30 more** agents (Windsurf, Cline, Roo Code, Codex, Continue, Augment, and others). See the [full list](https://getpromptscript.dev/formatters/).
139
136
 
140
137
  ## Configuration
141
138
 
@@ -302,11 +299,10 @@ docker run --rm -v $(pwd):/workspace ghcr.io/mrwogu/promptscript:latest validate
302
299
 
303
300
  ## Documentation
304
301
 
305
- - [Getting Started](https://getpromptscript.dev/getting-started/) 5-minute quickstart
306
- - [Language Reference](https://getpromptscript.dev/reference/syntax/) full syntax docs
307
- - [Inheritance Guide](https://getpromptscript.dev/guides/inheritance/) composition patterns
308
- - [Migration Guide](https://getpromptscript.dev/guides/migration/) converting existing files
309
- - [Enterprise Guide](https://getpromptscript.dev/guides/enterprise/) — scaling across organizations
302
+ - [Getting Started](https://getpromptscript.dev/getting-started/) - 5-minute quickstart
303
+ - [Language Reference](https://getpromptscript.dev/reference/syntax/) - full syntax docs
304
+ - [Guides](https://getpromptscript.dev/guides/) - inheritance, registry, migration, and more
305
+ - [Enterprise](https://getpromptscript.dev/guides/enterprise/) - scaling across organizations
310
306
 
311
307
  ## License
312
308