Skip to content

Build AI methods with your coding agent, then run them anywhere: as an MCP for chatbots, as a webapp for people, or via API for your software.

Feature/new plugin readme (#843)

Louis Choqueld89f20df6a91
diff --git a/README.md b/README.md
index cc78f7701..427fa5f1f 100644
--- a/README.md
+++ b/README.md
@@ -63,7 +63,7 @@ From here, Pipelex handles model routing across 60+ models, structured output pa
Paste this into Claude Code:
```
-Install mthds: run bash -c 'npm install -g mthds && mthds-agent bootstrap && claude plugin marketplace add mthds-ai/mthds-plugins && claude plugin install mthds@mthds-plugins'. Then tell me to restart Claude Code.
+Install mthds: run bash -c 'npm install -g mthds && mthds-agent bootstrap && claude plugin marketplace add mthds-ai/mthds-plugins && claude plugin install mthds@mthds-plugins'. Then tell me to /reload-plugins.
```
## With Codex
@@ -71,7 +71,7 @@ Install mthds: run bash -c 'npm install -g mthds && mthds-agent bootstrap && cla
Paste this into Codex:
```
-Install mthds: run bash -c 'npm install -g mthds && mthds-agent bootstrap && bash <(curl -fsSL https://raw.githubusercontent.com/mthds-ai/mthds-plugins/main/bin/install-codex.sh)'. Then tell me to run /plugins, search for MTHDS, and install it.
+Install mthds: run bash -c 'npm install -g mthds && mthds-agent bootstrap && mthds-agent codex install-hook && mthds-agent codex apply-config && codex plugin marketplace add mthds-ai/mthds-plugins'. Then tell me to restart Codex and run /plugins to install mthds.
```
Install the [VS Code extension](https://go.pipelex.com/vscode) for `.mthds` syntax highlighting and flowchart visualization.
diff --git a/pipelex/builder/operations/pipe_ops.py b/pipelex/builder/operations/pipe_ops.py
index 31c8b82a8..3a29768ba 100644
--- a/pipelex/builder/operations/pipe_ops.py
+++ b/pipelex/builder/operations/pipe_ops.py
@@ -30,6 +30,9 @@ _PIPE_CODE_ALIASES = ("pipe", "the_pipe_code", "code", "name", "pipe_name", "pip
# Aliases that agents may use instead of "output". First found is promoted when canonical key is absent; extras are dropped.
_OUTPUT_ALIASES = ("output_concept", "output_type")
+# Aliases that agents may use instead of "prompt". First found is promoted when canonical key is absent; extras are dropped.
+_PROMPT_ALIASES = ("prompt_template",)
+
def _normalize_sub_pipe_dict(data: dict[str, Any]) -> None:
"""Normalize a step/branch dict: resolve pipe_code aliases and drop extraneous fields."""
@@ -49,6 +52,16 @@ def _normalize_pipe_code_aliases(data: dict[str, Any]) -> None:
data.pop(alias)
+def _normalize_prompt_aliases(data: dict[str, Any]) -> None:
+ """Convert any alias of ``prompt`` to the canonical field name, in-place."""
+ for alias in _PROMPT_ALIASES:
+ if alias in data:
+ if "prompt" not in data:
+ data["prompt"] = data.pop(alias)
+ else:
+ data.pop(alias)
+
+
def parse_pipe_spec(pipe_type: str, spec_data: dict[str, Any]) -> PipeSpec:
"""Parse and validate a PipeSpec from JSON-like data.
@@ -79,6 +92,9 @@ def parse_pipe_spec(pipe_type: str, spec_data: dict[str, Any]) -> PipeSpec:
# Accept common aliases for "pipe_code" at the top level
_normalize_pipe_code_aliases(spec_data)
+ # Accept common aliases for "prompt" (e.g. "prompt_template")
+ _normalize_prompt_aliases(spec_data)
+
# Handle steps/branches conversion — normalize aliases and drop unknown fields.
# Deep-copy nested dicts to avoid mutating caller's nested structures.
if "steps" in spec_data:
diff --git a/tests/unit/pipelex/builder/operations/test_parse_pipe_spec.py b/tests/unit/pipelex/builder/operations/test_parse_pipe_spec.py
index 07cc61a11..bb4de57bc 100644
--- a/tests/unit/pipelex/builder/operations/test_parse_pipe_spec.py
+++ b/tests/unit/pipelex/builder/operations/test_parse_pipe_spec.py
@@ -113,6 +113,36 @@ class TestParsePipeSpec:
result = parse_pipe_spec("PipeLLM", spec)
assert result.output == "FirstAlias"
+ # -- prompt aliases ---------------------------------------------------
+
+ def test_prompt_template_alias_for_llm(self) -> None:
+ """`prompt_template` alone is promoted to `prompt` for PipeLLM."""
+ spec = {key: val for key, val in _BASE_LLM.items() if key != "prompt"}
+ spec["prompt_template"] = "Write about @text"
+ result = parse_pipe_spec("PipeLLM", spec)
+ assert isinstance(result, PipeLLMSpec)
+ assert result.prompt == "Write about @text"
+
+ def test_canonical_prompt_takes_precedence_over_alias(self) -> None:
+ """When both `prompt` and `prompt_template` are present, canonical wins and alias is dropped."""
+ spec = {**_BASE_LLM, "prompt_template": "alias_value"}
+ result = parse_pipe_spec("PipeLLM", spec)
+ assert isinstance(result, PipeLLMSpec)
+ assert result.prompt == _BASE_LLM["prompt"]
+
+ def test_prompt_template_alias_for_img_gen(self) -> None:
+ """`prompt_template` alias also works for PipeImgGen."""
+ spec: dict[str, Any] = {
+ "pipe_code": "my_img",
+ "description": "Generate image",
+ "inputs": {"prompt_text": "ImgGenPrompt"},
+ "output": "Image",
+ "prompt_template": "Generate: $prompt_text",
+ }
+ result = parse_pipe_spec("PipeImgGen", spec)
+ assert isinstance(result, PipeImgGenSpec)
+ assert result.prompt == "Generate: $prompt_text"
+
# -- steps/branches 'pipe' → 'pipe_code' alias -----------------------
def test_sequence_steps_canonical_pipe_code(self) -> None: