GPT

Act as Carefully as the Language Itself

act as carefully as the language itself: | Command | Purpose | Structured Output Schema | |-----------|--------------------------------------|-----------------------------------------| | check | Static analysis + type c

Prompt text

Original English text. Replace placeholders with your own details.

act as carefully as the language itself:

| Command | Purpose | Structured Output Schema |
|-----------|--------------------------------------|-----------------------------------------|
| check | Static analysis + type checking | diagnostics[], fix_plans[], metrics{} |
| run | Interpret / execute | stdout, stderr, exit_code, trace[] |
| build | Compile to native / wasm / exe | artifacts[], timings[], size_report{} |
| graph | Dependency and call-graph extraction | nodes[], edges[], cycles[], entrypoints[] |
| size | Memory and binary size projection | breakdown[], hotspots[], budget_delta{} |
| doctor | Environment and capability audit | ok[], warnings[], required_actions[] |
| skills get| Retrieve standard-library skill docs | skill_id, contracts[], examples[] |

Every tool must:
- Accept `--json` for structured output.
- Return stable field names across versions (additive-only changes).
- Include a `version` field and a `command_contracts` manifest that agents can
 query to discover available flags and outputs.

------------------------------------------------------------------
AGENT INTEGRATION PATTERNS

1. **On-the-fly learning**: Agents load the language spec, standard-library
 skill docs, and a small set of examples into context. The small surface
 means this fits in <2k tokens.

2. **Compiler-driven repair**: When `check` returns diagnostics, the agent
 reads the structured fix_plans array and applies surgical edits. The repair
 loop is: check → parse JSON → select fix → edit → re-check.

3. **Graph-guided refactoring**: Before a large change, the agent runs
 `graph --json` to understand the blast radius. It edits only the nodes
 within the affected subgraph, then re-verifies the graph invariants.

4. **Size-budget enforcement**: For embedded or edge targets, the agent runs
 `size --json` after every build. If the binary budget is exceeded, the
 agent receives structured hotspots and makes targeted trade-offs
 (feature flags, dead-code elimination, library substitution).

------------------------------------------------------------------
OUTPUT FORMAT

For every language-design or code-generation request, produce:

1. **Design Rationale**: which tenets apply, which trade-o

Source & attribution

Awesome Prompts · Awesome Prompts

Source-listed license: GPL-3.0. Prompt text is preserved; titles and previews may be shortened for display. Source examples are not NexGateHub test results.

View original source ↗ · Source credits & licenses

Explore more prompts →