Pencel

Appearance

Utilities ​

Pencel includes three utility integrations that handle foundational tasks: reading and writing local files, persisting knowledge, and searching the web. These are the building blocks that make your agents more capable across every workflow.

Filesystem ​

Trust LevelVerified
TransportBuiltin
Package@modelcontextprotocol/server-filesystem
ActivationAuto-connect recommended

What It Does ​

The Filesystem integration gives your agents the ability to read and write files on your local machine. This is how agents save reports, read configuration files, and work with data stored in your workspace.

  • Read files — open and read any file within the allowed directory.
  • Write files — create new files or overwrite existing ones.
  • List directories — browse folder contents.
  • Create directories — set up new folder structures.
  • Move and rename files and folders.
  • Search files by name pattern.

Setup ​

  1. Open Settings > Connections.
  2. Click Add Connection and select Filesystem from the catalog.
  3. Enter the directory path you want your agents to access (e.g., /Users/you/workspace).
  4. Click Connect.

WARNING

Your agent can read and write any file within the directory you specify. Do not point it at your entire home directory or system folders. Create a dedicated workspace folder and use that.

Example: Organize a Downloads Folder ​

List all files in /Users/me/workspace/incoming. Group them by file extension. Move PDFs to /reports, CSVs to /data, and images to /assets. Create a summary file listing what was moved where.

Example: Generate a Config File ​

Read the template at /templates/config-template.yaml. Replace the placeholder values with the settings from /data/environment.json. Save the result as /output/production-config.yaml.

Memory ​

Trust LevelVerified
TransportBuiltin
Package@modelcontextprotocol/server-memory
ActivationAuto-connect recommended

What It Does ​

The Memory integration gives your agents a persistent knowledge graph — a structured store of facts, relationships, and context that survives across sessions. Unlike conversation history (which is tied to a single chat), memory persists indefinitely.

  • Store facts — save key pieces of information as entities with properties.
  • Create relationships — link entities together (e.g., "Alice manages the Platform team").
  • Query knowledge — search for stored facts by entity or relationship type.
  • Update and delete — keep the knowledge graph accurate over time.

Setup ​

  1. Open Settings > Connections.
  2. Click Add Connection and select Memory from the catalog.
  3. Click Connect. No credentials needed.

The knowledge graph is stored locally on your machine.

TIP

Memory is especially powerful when combined with guidelines. Create a guideline that instructs your agents to save important facts to memory during runs. Over time, your agents build up institutional knowledge.

Example: Build a Team Directory ​

I just onboarded three new team members. Save these to memory: "Sarah Chen — Senior Engineer, Platform team, started March 2026." "Marcus Johnson — Product Manager, Growth team, started February 2026." "Priya Patel — Designer, Platform team, started March 2026." Also record that Sarah and Priya are both on the Platform team.

Later, any agent can query: "Who is on the Platform team?" and get the answer from memory.

Example: Track Project Decisions ​

Save to memory: "We decided on March 1st to use PostgreSQL instead of MongoDB for the analytics service. The reason was better support for complex joins and lower operational cost. Decision made by the Platform team."

This creates a searchable record of decisions that any agent can reference.

Trust LevelVerified
TransportBuiltin
Package@modelcontextprotocol/server-brave-search
ActivationWhen relevant recommended

What It Does ​

Brave Search gives your agents the ability to search the web for current information. This is essential for tasks that require up-to-date data — market research, competitor analysis, industry news, or fact-checking.

  • Web search — search the open web and get ranked results.
  • News search — find recent news articles on a topic.
  • Result summaries — get titles, URLs, and snippets for each result.

Setup ​

Step 1: Get a Brave Search API Key

  1. Go to brave.com/search/api and create an account.
  2. Subscribe to the Free plan (2,000 queries per month) or a paid plan.
  3. Copy your API key (starts with BSA).

Step 2: Add the Connection

  1. Open Settings > Connections.
  2. Click Add Connection and select Brave Search from the catalog.
  3. Paste your API key into the credential field.
  4. Click Connect.

TIP

The free plan (2,000 queries per month) is enough for most automation workflows. Upgrade only if you run search-heavy workflows frequently.

Example: Market Research ​

Search the web for "enterprise AI automation tools 2026." Compile a list of the top 10 competitors mentioned in the results. For each, note their key features, pricing model, and target market. Save the research to /reports/market-landscape-[date].md.

Example: Industry News Digest ​

Search for news about "supply chain disruption" from the last 7 days. Summarize the top 5 articles in 2-3 sentences each, including the source and date. Post the digest to the #industry-news Slack channel.