Add runapi-mcp plugin#568
Conversation
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (4)
✅ Files skipped from review due to trivial changes (1)
🚧 Files skipped from review as they are similar to previous changes (3)
📝 WalkthroughSummary by CodeRabbit
WalkthroughThis PR adds the complete RunAPI MCP plugin integration, including plugin metadata versioning, MCP command configuration, comprehensive documentation, agent orchestration prompts for model discovery and task execution, and user-facing command flows for balance checking, model search, API setup, and quick media generation. ChangesRunAPI MCP Plugin
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Suggested reviewers
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
✨ Simplify code
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 4
🧹 Nitpick comments (7)
plugins/runapi-mcp/agents/task-executor.md (2)
21-23: ⚖️ Poor tradeoffConsider using action verbs instead of explicit tool names.
The guidelines recommend using action verbs in the markdown body rather than tool-name vocabulary. While the adapter rewrites a conservative set, unrewritten cases may bleed through to hosts like Codex or Cursor.
Consider rephrasing instructions to use action-oriented language where feasible.
As per coding guidelines: Use action verbs in markdown body, not tool-name vocabulary: prefer "open the file" over "use the Read tool". The adapter rewrites a conservative set, but unrewritten cases may bleed through.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@plugins/runapi-mcp/agents/task-executor.md` around lines 21 - 23, The markdown uses explicit tool names (mcp__runapi__create_task, mcp__runapi__get_task) instead of action verbs; replace those occurrences with action-oriented instructions (e.g., "create a task with the provided values" and "retrieve/check a task by ID") and keep the same semantic behavior and return requirements (compact response: task ID, status, output URLs, cost fields) so the adapter can rewrite tool references safely while preserving the calls to mcp__runapi__create_task and mcp__runapi__get_task under the hood.
2-4: 💤 Low valueAlign description with "Use when…" pattern.
The coding guidelines specify that the agent description should include a "Use when…" trigger. Consider starting the description with "Use when" for consistency with the documented pattern.
📝 Optional rephrasing
description: >- - Media task execution agent. Delegates here for create_task calls so the - main conversation stays focused. Spawn one per task for parallel generation. + Use when creating or checking RunAPI media tasks while keeping the main + conversation focused. Spawn one per task for parallel generation.Based on learnings: Agents frontmatter must include:
name,description(with "Use when…" trigger),model: opus|sonnet|haiku|inherit, optionaltools:, optionalcolor:.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@plugins/runapi-mcp/agents/task-executor.md` around lines 2 - 4, Update the frontmatter in task-executor.md so the description field begins with the required "Use when…" trigger (e.g., "Use when delegating media task execution for create_task calls; spawn one per task for parallel generation.") and preserve the existing purpose text; ensure the agent frontmatter also includes the required fields (name and model set to one of opus|sonnet|haiku|inherit) and any optional tools or color entries so the metadata remains valid.plugins/runapi-mcp/commands/balance.md (1)
8-8: ⚖️ Poor tradeoffConsider using action verbs instead of explicit tool names.
The guidelines recommend using action verbs in the markdown body rather than tool-name vocabulary. While the adapter rewrites a conservative set, unrewritten cases may bleed through to hosts like Codex or Cursor.
As per coding guidelines: Use action verbs in markdown body, not tool-name vocabulary: prefer "open the file" over "use the Read tool". The adapter rewrites a conservative set, but unrewritten cases may bleed through.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@plugins/runapi-mcp/commands/balance.md` at line 8, Replace the explicit tool-name phrasing "Call `mcp__runapi__check_balance`" with an action-verb description that avoids exposing the tool identifier directly; for example, change the sentence to "Invoke the runapi check balance action" or "Execute the runapi check balance operation" while keeping reference to the underlying identifier mcp__runapi__check_balance only in code blocks or metadata if needed (update the text around the symbol in the commands/balance.md content so the markdown body uses an action verb instead of the raw tool name).plugins/runapi-mcp/commands/models.md (1)
14-17: ⚖️ Poor tradeoffConsider using action verbs instead of explicit tool names.
The guidelines recommend using action verbs in the markdown body rather than tool-name vocabulary. While the adapter rewrites a conservative set, unrewritten cases may bleed through to hosts like Codex or Cursor.
As per coding guidelines: Use action verbs in markdown body, not tool-name vocabulary: prefer "open the file" over "use the Read tool". The adapter rewrites a conservative set, but unrewritten cases may bleed through.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@plugins/runapi-mcp/commands/models.md` around lines 14 - 17, Replace explicit tool-name vocabulary in the markdown items by using action verbs and generic descriptions instead of raw tool identifiers: e.g., reword the line that mentions mcp__runapi__list_models to "list models with a narrow filter," change the mcp__runapi__get_model_info reference to "retrieve detailed info for a single model (include service and action when known)," and change mcp__runapi__check_pricing to "check pricing for the model" — update items 2–5 to use these action-verb phrasings while keeping the same semantics.plugins/runapi-mcp/agents/model-advisor.md (2)
2-4: 💤 Low valueAlign description with "Use when…" pattern.
The coding guidelines specify that the agent description should include a "Use when…" trigger. While the current description mentions "Delegates here when the user needs...", consider starting with "Use when" for consistency with the documented pattern.
📝 Optional rephrasing
description: >- - Model discovery and recommendation agent. Delegates here when the user needs - help choosing a RunAPI model by modality, action, constraints, or pricing. + Use when the user needs help choosing a RunAPI model by modality, action, + constraints, or pricing. Handles model discovery and recommendations.Based on learnings: Agents frontmatter must include:
name,description(with "Use when…" trigger),model: opus|sonnet|haiku|inherit, optionaltools:, optionalcolor:.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@plugins/runapi-mcp/agents/model-advisor.md` around lines 2 - 4, Update the frontmatter description to follow the "Use when…" pattern by replacing the current sentence that begins "Delegates here when the user needs..." with a leading "Use when" trigger (e.g., "Use when the user needs help choosing a RunAPI model by modality, action, constraints, or pricing."). Ensure the frontmatter includes the required keys (name and model) and that the description field explicitly begins with "Use when" and concisely describes the triggering conditions; keep optional keys (tools, color) as needed.
19-23: ⚖️ Poor tradeoffConsider using action verbs instead of explicit tool names.
The guidelines recommend using action verbs in the markdown body rather than tool-name vocabulary (e.g., "open the file" instead of "use the Read tool"). While the adapter rewrites a conservative set, unrewritten cases may bleed through to hosts like Codex or Cursor.
Consider rephrasing instructions to use action-oriented language where feasible, though explicit tool names may be acceptable if the adapter handles them.
As per coding guidelines: Use action verbs in markdown body, not tool-name vocabulary: prefer "open the file" over "use the Read tool". The adapter rewrites a conservative set, but unrewritten cases may bleed through.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@plugins/runapi-mcp/agents/model-advisor.md` around lines 19 - 23, The markdown uses explicit tool names (e.g., `mcp__runapi__list_models`, `mcp__runapi__get_model_info`, `mcp__runapi__check_pricing`, and `create_task`) instead of action-oriented phrasing; update the steps to use action verbs in the body while keeping the precise tool identifiers for implementation/reference only (for example, reword "Call `mcp__runapi__list_models`..." to "List models with the narrowest useful modality/service/action filter" and similarly change "call `mcp__runapi__get_model_info`" to "retrieve model info for promising candidates" and "call `mcp__runapi__check_pricing`" to "check pricing when cost matters"), and ensure the final step still provides the exact service, action, and model slug required for create_task.plugins/runapi-mcp/commands/gen.md (1)
15-16: ⚡ Quick winPrefer action verbs over explicit tool names.
Lines 15-16 and 22 directly reference MCP tool names (
mcp__runapi__list_models,mcp__runapi__get_model_info,mcp__runapi__check_pricing). Per the coding guidelines, command instructions should use action verbs instead of tool-name vocabulary to ensure better cross-harness transpilation.♻️ Proposed refactor
1. If `$ARGUMENTS` is empty, ask for the prompt or task description. -2. Call `mcp__runapi__list_models` with the most likely modality filter. -3. Call `mcp__runapi__get_model_info` with the selected service, action, and model slug, then obey returned input rules. +2. List RunAPI models with the most likely modality filter. +3. Get model info for the selected service, action, and model slug, then obey returned input rules. 4. If the selected task is video, music, or a batch, ask for confirmation before creating it.Do not invent model slugs. -Do not quote prices without `mcp__runapi__check_pricing`. +Do not quote prices without checking pricing first. Do not describe generated media content.As per coding guidelines: Use action verbs in markdown body, not tool-name vocabulary: prefer "open the file" over "use the Read tool". The adapter rewrites a conservative set, but unrewritten cases may bleed through
Also applies to: 22-22
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@plugins/runapi-mcp/commands/gen.md` around lines 15 - 16, The markdown uses explicit MCP tool names (mcp__runapi__list_models, mcp__runapi__get_model_info, mcp__runapi__check_pricing); update the phrasing to use action verbs instead (e.g., "list models filtered by modality", "retrieve model info for the selected service/action/slug", "check model pricing") so instructions are tool-agnostic and follow the coding guideline — replace occurrences of those exact identifiers in the text with concise action-verb alternatives while preserving the same parameters and intent.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@plugins/runapi-mcp/agents/model-advisor.md`:
- Around line 1-6: The agent frontmatter in model-advisor.md is missing the
required "name" field; add a top-level name entry to the YAML frontmatter (e.g.,
name: "Model Advisor") and ensure the description includes the required "Use
when…" trigger phrase, keep model set to one of the allowed values (model:
haiku|sonnet|opus|inherit), and optionally include tools: and color: keys if
applicable; update the frontmatter block containing description and model so it
includes name, description (with "Use when…"), model, and any optional
tools/color entries.
In `@plugins/runapi-mcp/agents/task-executor.md`:
- Around line 1-7: Add a required name entry to the YAML frontmatter of
plugins/runapi-mcp/agents/task-executor.md and update the description to include
the required "Use when…" trigger; specifically, in the existing frontmatter that
currently has description, model: inherit and tools, add a top-level name:
"task-executor" (or another concise identifier), change description to start
with a short summary then include a "Use when…" sentence, keep model: inherit
and tools: mcp__runapi__create_task, mcp__runapi__get_task as-is, and optionally
add a color: field if desired so the frontmatter meets the agent guidelines.
In `@plugins/runapi-mcp/commands/balance.md`:
- Around line 1-4: The command frontmatter is missing the required argument-hint
field; update the YAML frontmatter that currently contains only description to
also include an argument-hint entry (e.g., argument-hint: "[none]" or
argument-hint: "") so the frontmatter now contains both description and
argument-hint as required by the guidelines.
In `@plugins/runapi-mcp/commands/setup.md`:
- Around line 1-5: The frontmatter for the RunAPI setup command is missing the
required argument-hint field; update the YAML frontmatter that currently
contains description to also include an argument-hint entry (e.g.,
argument-hint: "" or a short placeholder) so it conforms to the commands
frontmatter requirement; modify the top-of-file frontmatter block that contains
description to add argument-hint.
---
Nitpick comments:
In `@plugins/runapi-mcp/agents/model-advisor.md`:
- Around line 2-4: Update the frontmatter description to follow the "Use when…"
pattern by replacing the current sentence that begins "Delegates here when the
user needs..." with a leading "Use when" trigger (e.g., "Use when the user needs
help choosing a RunAPI model by modality, action, constraints, or pricing.").
Ensure the frontmatter includes the required keys (name and model) and that the
description field explicitly begins with "Use when" and concisely describes the
triggering conditions; keep optional keys (tools, color) as needed.
- Around line 19-23: The markdown uses explicit tool names (e.g.,
`mcp__runapi__list_models`, `mcp__runapi__get_model_info`,
`mcp__runapi__check_pricing`, and `create_task`) instead of action-oriented
phrasing; update the steps to use action verbs in the body while keeping the
precise tool identifiers for implementation/reference only (for example, reword
"Call `mcp__runapi__list_models`..." to "List models with the narrowest useful
modality/service/action filter" and similarly change "call
`mcp__runapi__get_model_info`" to "retrieve model info for promising candidates"
and "call `mcp__runapi__check_pricing`" to "check pricing when cost matters"),
and ensure the final step still provides the exact service, action, and model
slug required for create_task.
In `@plugins/runapi-mcp/agents/task-executor.md`:
- Around line 21-23: The markdown uses explicit tool names
(mcp__runapi__create_task, mcp__runapi__get_task) instead of action verbs;
replace those occurrences with action-oriented instructions (e.g., "create a
task with the provided values" and "retrieve/check a task by ID") and keep the
same semantic behavior and return requirements (compact response: task ID,
status, output URLs, cost fields) so the adapter can rewrite tool references
safely while preserving the calls to mcp__runapi__create_task and
mcp__runapi__get_task under the hood.
- Around line 2-4: Update the frontmatter in task-executor.md so the description
field begins with the required "Use when…" trigger (e.g., "Use when delegating
media task execution for create_task calls; spawn one per task for parallel
generation.") and preserve the existing purpose text; ensure the agent
frontmatter also includes the required fields (name and model set to one of
opus|sonnet|haiku|inherit) and any optional tools or color entries so the
metadata remains valid.
In `@plugins/runapi-mcp/commands/balance.md`:
- Line 8: Replace the explicit tool-name phrasing "Call
`mcp__runapi__check_balance`" with an action-verb description that avoids
exposing the tool identifier directly; for example, change the sentence to
"Invoke the runapi check balance action" or "Execute the runapi check balance
operation" while keeping reference to the underlying identifier
mcp__runapi__check_balance only in code blocks or metadata if needed (update the
text around the symbol in the commands/balance.md content so the markdown body
uses an action verb instead of the raw tool name).
In `@plugins/runapi-mcp/commands/gen.md`:
- Around line 15-16: The markdown uses explicit MCP tool names
(mcp__runapi__list_models, mcp__runapi__get_model_info,
mcp__runapi__check_pricing); update the phrasing to use action verbs instead
(e.g., "list models filtered by modality", "retrieve model info for the selected
service/action/slug", "check model pricing") so instructions are tool-agnostic
and follow the coding guideline — replace occurrences of those exact identifiers
in the text with concise action-verb alternatives while preserving the same
parameters and intent.
In `@plugins/runapi-mcp/commands/models.md`:
- Around line 14-17: Replace explicit tool-name vocabulary in the markdown items
by using action verbs and generic descriptions instead of raw tool identifiers:
e.g., reword the line that mentions mcp__runapi__list_models to "list models
with a narrow filter," change the mcp__runapi__get_model_info reference to
"retrieve detailed info for a single model (include service and action when
known)," and change mcp__runapi__check_pricing to "check pricing for the model"
— update items 2–5 to use these action-verb phrasings while keeping the same
semantics.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro Plus
Run ID: 40778de0-7016-422a-91e9-951942669854
📒 Files selected for processing (10)
plugins/runapi-mcp/.claude-plugin/plugin.jsonplugins/runapi-mcp/.codex-plugin/plugin.jsonplugins/runapi-mcp/.mcp.jsonplugins/runapi-mcp/README.mdplugins/runapi-mcp/agents/model-advisor.mdplugins/runapi-mcp/agents/task-executor.mdplugins/runapi-mcp/commands/balance.mdplugins/runapi-mcp/commands/gen.mdplugins/runapi-mcp/commands/models.mdplugins/runapi-mcp/commands/setup.md
📜 Review details
🧰 Additional context used
📓 Path-based instructions (4)
plugins/**/{.claude-plugin/plugin.json,agents/*.md,skills/*/SKILL.md,commands/*.md}
📄 CodeRabbit inference engine (.cursor/rules/000-project-conventions.mdc)
Plugin files follow the structure:
plugins/<name>/{.claude-plugin/plugin.json, agents/*.md, skills/<name>/SKILL.md, commands/*.md}and are auto-discovered
Files:
plugins/runapi-mcp/agents/task-executor.mdplugins/runapi-mcp/commands/balance.mdplugins/runapi-mcp/commands/models.mdplugins/runapi-mcp/.claude-plugin/plugin.jsonplugins/runapi-mcp/commands/setup.mdplugins/runapi-mcp/commands/gen.mdplugins/runapi-mcp/agents/model-advisor.md
plugins/*/agents/*.md
📄 CodeRabbit inference engine (.cursor/rules/020-agent-skill-authoring.mdc)
plugins/*/agents/*.md: Agents frontmatter must include:name,description(with "Use when…" trigger),model: opus|sonnet|haiku|inherit, optionaltools:, optionalcolor:
Do not name agentsdefault,worker, orexplorer— they collide with Codex built-insSubagent files must follow format
plugins/*/agents/<name>.mdfor cross-harness transpilation
Files:
plugins/runapi-mcp/agents/task-executor.mdplugins/runapi-mcp/agents/model-advisor.md
{plugins/*/agents/*.md,plugins/*/skills/*/SKILL.md,plugins/*/commands/*.md}
📄 CodeRabbit inference engine (.cursor/rules/020-agent-skill-authoring.mdc)
{plugins/*/agents/*.md,plugins/*/skills/*/SKILL.md,plugins/*/commands/*.md}: Use action verbs in markdown body, not tool-name vocabulary: prefer "open the file" over "use the Read tool". The adapter rewrites a conservative set, but unrewritten cases may bleed through
Do not referenceTodoWrite, theTaskspawn tool, or hooks as load-bearing — they don't exist in Codex or Cursor
Files:
plugins/runapi-mcp/agents/task-executor.mdplugins/runapi-mcp/commands/balance.mdplugins/runapi-mcp/commands/models.mdplugins/runapi-mcp/commands/setup.mdplugins/runapi-mcp/commands/gen.mdplugins/runapi-mcp/agents/model-advisor.md
plugins/*/commands/*.md
📄 CodeRabbit inference engine (.cursor/rules/020-agent-skill-authoring.mdc)
Commands frontmatter must include:
description,argument-hint
Files:
plugins/runapi-mcp/commands/balance.mdplugins/runapi-mcp/commands/models.mdplugins/runapi-mcp/commands/setup.mdplugins/runapi-mcp/commands/gen.md
🧠 Learnings (8)
📓 Common learnings
Learnt from: CR
Repo: wshobson/agents PR: 0
File: .cursor/rules/020-agent-skill-authoring.mdc:0-0
Timestamp: 2026-05-29T19:12:43.515Z
Learning: Applies to plugins/*/agents/*.md : Agents frontmatter must include: `name`, `description` (with "Use when…" trigger), `model: opus|sonnet|haiku|inherit`, optional `tools:`, optional `color:`
📚 Learning: 2026-05-11T13:52:18.563Z
Learnt from: jau123
Repo: wshobson/agents PR: 0
File: :0-0
Timestamp: 2026-05-11T13:52:18.563Z
Learning: In the wshobson/agents repository, all plugin manifest files at `plugins/*/.claude-plugin/plugin.json` follow a consistent 5-field schema: `name`, `version`, `description`, `author`, and `license`. Suggestions to remove any of these fields diverge from the established repository convention. Version-bump PRs should only update the `version` field.
Applied to files:
plugins/runapi-mcp/.codex-plugin/plugin.jsonplugins/runapi-mcp/.claude-plugin/plugin.json
📚 Learning: 2026-05-29T19:12:32.497Z
Learnt from: CR
Repo: wshobson/agents PR: 0
File: .cursor/rules/000-project-conventions.mdc:0-0
Timestamp: 2026-05-29T19:12:32.497Z
Learning: Applies to .claude-plugin/marketplace.json : Never modify `.claude-plugin/marketplace.json` from inside a plugin PR — it's the registry
Applied to files:
plugins/runapi-mcp/.claude-plugin/plugin.json
📚 Learning: 2026-05-29T19:12:43.515Z
Learnt from: CR
Repo: wshobson/agents PR: 0
File: .cursor/rules/020-agent-skill-authoring.mdc:0-0
Timestamp: 2026-05-29T19:12:43.515Z
Learning: Applies to plugins/*/commands/*.md : Commands frontmatter must include: `description`, `argument-hint`
Applied to files:
plugins/runapi-mcp/commands/gen.md
📚 Learning: 2026-05-29T19:12:32.497Z
Learnt from: CR
Repo: wshobson/agents PR: 0
File: .cursor/rules/000-project-conventions.mdc:0-0
Timestamp: 2026-05-29T19:12:32.497Z
Learning: Applies to commands/** : Do not hand-edit anything under `commands/` — regenerate via `make generate HARNESS=<x>`
Applied to files:
plugins/runapi-mcp/commands/gen.md
📚 Learning: 2026-05-29T19:12:43.515Z
Learnt from: CR
Repo: wshobson/agents PR: 0
File: .cursor/rules/020-agent-skill-authoring.mdc:0-0
Timestamp: 2026-05-29T19:12:43.515Z
Learning: Applies to plugins/*/agents/*.md : Agents frontmatter must include: `name`, `description` (with "Use when…" trigger), `model: opus|sonnet|haiku|inherit`, optional `tools:`, optional `color:`
Applied to files:
plugins/runapi-mcp/agents/model-advisor.md
📚 Learning: 2026-05-29T19:12:23.170Z
Learnt from: CR
Repo: wshobson/agents PR: 0
File: GEMINI.md:0-0
Timestamp: 2026-05-29T19:12:23.170Z
Learning: Applies to agents/**/*.md : Model assignment can be overridden at the per-agent level via `model:` frontmatter
Applied to files:
plugins/runapi-mcp/agents/model-advisor.md
📚 Learning: 2026-05-29T19:12:14.231Z
Learnt from: CR
Repo: wshobson/agents PR: 0
File: CLAUDE.md:0-0
Timestamp: 2026-05-29T19:12:14.231Z
Learning: Follow agent design patterns and conventions documented in AGENTS.md
Applied to files:
plugins/runapi-mcp/agents/model-advisor.md
🔇 Additional comments (5)
plugins/runapi-mcp/.claude-plugin/plugin.json (1)
1-9: LGTM!plugins/runapi-mcp/.mcp.json (1)
1-9: LGTM!plugins/runapi-mcp/README.md (1)
1-56: LGTM!plugins/runapi-mcp/commands/models.md (1)
1-5: LGTM!plugins/runapi-mcp/.codex-plugin/plugin.json (1)
2-2:⚠️ Potential issue | 🟡 Minor | ⚡ Quick winVerify name consistency across plugin manifests.
The Claude plugin manifest uses
"name": "runapi"while this Codex manifest uses"name": "runapi-mcp". If both hosts rely on the name field for plugin identity or cross-references, this inconsistency could cause discovery or configuration issues.Run the following script to check if other plugins follow a consistent naming pattern across Claude and Codex manifests:
#!/bin/bash # Description: Compare plugin names across Claude and Codex manifests to identify naming patterns fd -t f 'plugin.json' plugins/ --exec sh -c ' for f; do dir=$(dirname "$f") plugin=$(echo "$dir" | cut -d/ -f2) host=$(basename "$(dirname "$f")") name=$(jq -r ".name" "$f") echo "$plugin|$host|$name" done ' sh {} + | sort⛔ Skipped due to learnings
Learnt from: jau123 Repo: wshobson/agents PR: 0 File: :0-0 Timestamp: 2026-05-11T13:52:18.563Z Learning: In the wshobson/agents repository, all plugin manifest files at `plugins/*/.claude-plugin/plugin.json` follow a consistent 5-field schema: `name`, `version`, `description`, `author`, and `license`. Suggestions to remove any of these fields diverge from the established repository convention. Version-bump PRs should only update the `version` field.Learnt from: CR Repo: wshobson/agents PR: 0 File: .cursor/rules/000-project-conventions.mdc:0-0 Timestamp: 2026-05-29T19:12:32.497Z Learning: Applies to plugins/**/{.claude-plugin/plugin.json,agents/*.md,skills/*/SKILL.md,commands/*.md} : Plugin files follow the structure: `plugins/<name>/{.claude-plugin/plugin.json, agents/*.md, skills/<name>/SKILL.md, commands/*.md}` and are auto-discoveredLearnt from: CR Repo: wshobson/agents PR: 0 File: AGENTS.md:0-0 Timestamp: 2026-06-02T23:10:39.225Z Learning: Applies to plugins/*/package.json : Plugin names must be lowercase and hyphen-separated, never use `__` (reserved for adapter namespace separator)
1 participant
Summary
Adds RunAPI MCP, a media generation plugin that gives Claude Code, Codex, and other MCP hosts access to 130+ AI models across 18 providers.
What it does
Models covered
Install
Links