perigon-automation
by ComposioHQperigon-automation helps Claude run Perigon tasks through Composio Rube MCP by discovering current tools, checking the Perigon connection, and executing with less schema guesswork.
This skill scores 68/100, which means it is acceptable for directory listing but should be presented as a lightweight connector guide rather than a fully worked automation package. It gives agents enough trigger and setup guidance to use Perigon through Rube MCP with less guesswork than a generic prompt, but directory users should expect to depend heavily on live tool discovery because the repository includes only a single SKILL.md and no concrete task implementations.
- Valid skill frontmatter clearly declares the `rube` MCP requirement and describes Perigon automation through Composio/Rube.
- Prerequisites and setup steps tell the agent to verify `RUBE_SEARCH_TOOLS`, manage a `perigon` connection, and confirm ACTIVE status before workflows.
- The repeated instruction to call `RUBE_SEARCH_TOOLS` first helps reduce schema drift and gives agents a concrete operational pattern rather than hardcoded stale tool calls.
- No bundled scripts, examples, references, or README beyond SKILL.md, so users must rely on live Rube tool discovery for exact Perigon operations and schemas.
- Install/setup guidance covers adding the Rube MCP endpoint and managing the Perigon connection, but there is no repository-level install command or tested end-to-end Perigon task example.
Overview of perigon-automation skill
What perigon-automation does
perigon-automation is a Claude skill for running Perigon operations through Composio’s Rube MCP interface. Its main job is not to hard-code one Perigon workflow; it gives the agent a safer pattern for discovering the current Perigon tool schema, checking authentication, and then executing the right Rube tool for the user’s news, media, or Perigon data task.
Best fit for Workflow Automation users
The perigon-automation skill is best for users who already use Claude with MCP and want workflow automation around Perigon without manually looking up Composio tool names each time. It fits teams that need repeatable Perigon actions, such as querying news data, enriching monitoring workflows, or connecting Perigon outputs to downstream research and reporting steps.
Key differentiator: search tools first
The most important behavior is the instruction to call RUBE_SEARCH_TOOLS before executing any workflow. This matters because Composio tool schemas can change, and Perigon operations may expose different required fields depending on the current toolkit. A generic prompt may guess tool names or stale parameters; this skill pushes the agent to discover current tool slugs, input schemas, recommended execution plans, and pitfalls before acting.
Adoption requirements and limits
Before installing or relying on perigon-automation, confirm your client supports MCP and can connect to Rube at https://rube.app/mcp. You also need an active Perigon connection through RUBE_MANAGE_CONNECTIONS with toolkit perigon. The repository contains a single SKILL.md and no helper scripts, rules, or reference files, so the value is in the operational pattern rather than a large packaged workflow library.
How to Use perigon-automation skill
perigon-automation install context
Install the perigon-automation skill from the Composio skills repository with:
npx skills add ComposioHQ/awesome-claude-skills --skill perigon-automation
Then configure Rube MCP in your AI client by adding https://rube.app/mcp as an MCP server. The upstream skill notes that no API key is needed for the MCP endpoint itself, but you still must authorize the Perigon toolkit connection when Rube asks for it.
First setup checks before running tasks
Start by confirming that RUBE_SEARCH_TOOLS is available. Next, ask the agent to run RUBE_MANAGE_CONNECTIONS with toolkits: ["perigon"]. If the returned status is not ACTIVE, complete the auth link that Rube provides and re-check status before requesting Perigon actions. Skipping this step is the most common reason a seemingly valid prompt fails.
A practical first prompt:
Use the perigon-automation skill. First verify that Rube MCP is available, then check whether the Perigon toolkit connection is ACTIVE. If it is not active, give me the auth step and stop. Do not run a Perigon operation until the connection is confirmed.
Turning a rough goal into a usable prompt
Weak prompt: “Get Perigon news about AI.”
Stronger prompt:
Use perigon-automation for Workflow Automation. Search current Rube tools for a Perigon task that retrieves recent English-language articles about enterprise AI funding. Show the tool you plan to use, required fields, and any schema assumptions. After confirming the Perigon connection is ACTIVE, run the tool for the last 7 days and summarize the results with source, date, headline, URL, and why each item is relevant.
This works better because it gives the agent a concrete use case, time range, language, output format, and a required discovery step. The skill is strongest when you describe the Perigon business task, not just the topic.
Repository files to read first
For this skill, read composio-skills/perigon-automation/SKILL.md first; it is effectively the full implementation. Pay attention to the sections on Prerequisites, Setup, Tool Discovery, and Core Workflow Pattern. There are no separate README.md, rules/, resources/, or scripts/ folders in the current file tree, so do not expect additional examples beyond the skill instructions.
perigon-automation skill FAQ
Is perigon-automation only for Perigon API experts?
No. The perigon-automation skill is useful for non-experts because it asks Rube to discover the available Perigon tools and schemas before execution. However, you still need to know the outcome you want: topic, date range, filters, fields, and how the result should be used. The skill can reduce schema guesswork, but it cannot infer a precise media intelligence workflow from a vague request.
How is it different from an ordinary Claude prompt?
An ordinary prompt may ask Claude to “use Perigon” and hope the model knows the right tool interface. perigon-automation adds a specific execution discipline: search Rube tools first, check the Perigon connection, inspect current schemas, then execute. That makes it better for live tool use where stale parameter names or unauthenticated connections would otherwise derail the workflow.
When should I not use this skill?
Do not use it if you do not have MCP support, cannot connect Rube, or do not plan to use the Perigon toolkit. It is also not a full news-analysis application by itself. If your goal is a custom dashboard, long-term storage, deduplication rules, or production alerting, treat this as the Perigon action layer and pair it with your own workflow orchestration.
Does the skill include reusable scripts or examples?
No. The current upstream package is a compact SKILL.md with instructions and example Rube calls, not a script bundle. That makes perigon-automation install lightweight, but it also means your team should create its own tested prompt templates for recurring Perigon workflows.
How to Improve perigon-automation skill
Improve perigon-automation inputs
The fastest way to improve results is to provide structured task details: target topic, geography, language, publication window, source type, required fields, maximum result count, and output format. For monitoring use cases, also state what should be excluded, such as press releases, duplicates, opinion pieces, or irrelevant company-name collisions.
Example add-on:
Exclude duplicate syndications, prioritize primary reporting, and flag any result where the company name may refer to a different organization.
Avoid common failure modes
The biggest failure modes are skipping RUBE_SEARCH_TOOLS, assuming a stale schema, running before the Perigon connection is ACTIVE, or asking for broad research without measurable filters. Make the agent state the discovered tool slug and required fields before execution. If a call fails, ask it to re-run tool discovery rather than inventing a parameter fix.
Iterate after the first output
After the first run, refine based on what the results reveal. If results are noisy, tighten entities, date ranges, or source types. If results are sparse, broaden keywords or remove overly strict filters. For recurring workflow automation, save the final prompt pattern that worked, including the discovery step and your preferred summary table.
What maintainers could add next
The skill would become more adoption-ready with a short README.md, two or three tested Perigon workflow examples, and troubleshooting notes for inactive Rube connections. A small prompt library for common Perigon usage—news search, source monitoring, entity tracking, and article enrichment—would make the perigon-automation guide more actionable without changing the core skill behavior.