Atlas / MCP servers / dave-london / Pare

PareSAFE

Name: Pare
Author: dave-london

mcp/dave-london/pare

MCP server by dave-london · ai coding · ai tools · claude · Audited 2026-10-07

⚑ Report / re-scan Source ↗

Dev tools, optimized for agents. Structured, token-efficient MCP servers for git, test runners, npm, Docker, and more.

Verdict

SAFE

Grade

Trust score

73 /100

Exposed tools

200 165r · 41w · 9d

Transport

stdio

License

MIT

Stars

138

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

Pare

[](https://github.com/Dave-London/Pare/actions/workflows/ci.yml) [](https://codecov.io/gh/Dave-London/Pare) [](https://www.npmjs.com/package/@paretools/git) [](https://www.npmjs.com/package/@paretools/git) [](https://www.typescriptlang.org/) [](https://github.com/Dave-London/Pare/blob/main/LICENSE) [](https://nodejs.org) [](https://scorecard.dev/viewer/?uri=github.com/Dave-London/Pare) [](https://www.bestpractices.dev/projects/11962)

Reliable, structured CLI output for AI agents — no more parsing fragile terminal text.

Pare provides MCP servers that wrap common developer tools (git, npm, docker, test runners, etc.) and return clean, schema-validated JSON instead of raw terminal text. Agents get typed data they can act on directly, without brittle string parsing.

The Problem

Parsing CLI output is fragile. Raw terminal text includes ANSI escape codes, decorative headers, progress bars, locale-specific formatting, and platform differences that break agent workflows in subtle ways. An agent that works fine with git status on macOS may fail on Windows because the output format changed. A test runner's summary line might shift between versions, silently breaking a regex.

Pare eliminates this entire

Read from source at commit 6f666d56ef4aOBSERVED · 2026-10-07

Connect

Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control.

claude-code (npm)

claude mcp add bazel -- npx -y @paretools/[email protected]

Exposed tools (200)

165 read · 41 write · 9 destructive. Blast radius: 9 tools can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.

Tool	Risk	Description
`a`	read	A
`add`	write
`add-package`	write
`ansible-galaxy`	read
`ansible-inventory`	read
`ansible-playbook`	read
`api`	read
`apply`	write
`archive`	read
`assemble`	read	Assembles the outputs of this project.
`audit`	read
`b`	read	B
`bazel`	read
`biome-check`	read
`biome-format`	read
`bisect`	read
`black`	read
`blame`	read
`branch`	read
`bug`	read	Something isn
`build`	read
`bundle-check`	read
`bundle-exec`	write
`bundle-install`	write
`c`	read	C
`check`	read
`checkout`	read
`cherry-pick`	read
`clean`	read
`clippy`	read
`cmake`	read
`commit`	write
`compose-build`	read
`compose-down`	read
`compose-logs`	read
`compose-ps`	read
`compose-up`	read
`conda`	read
`config`	read
`count`	read
`coverage`	read
`describe`	read
`develop`	read
`diff`	read
`discover-tools`	read
`discussion-list`	read
`doc`	read
`enhancement`	read	New feature
`env`	read
`esbuild`	read
`exec`	write
`feature`	read
`find`	read
`flake-check`	read
`flake-show`	read
`flake-update`	write
`fmt`	read
`format-check`	read
`gem-install`	write
`gem-list`	read
`gem-outdated`	read
`generate`	read
`get`	read
`gist-create`	write
`gitleaks`	read
`golangci-lint`	read
`gradle-build`	read
`gradle-dependencies`	read
`gradle-tasks`	read
`gradle-test`	read
`hadolint`	read
`head`	read
`helm`	read
`help`	read	Shows help.
`images`	read
`info`	read	Runs
`init`	read	Initializes a Terraform working directory. Downloads providers, configures backend, and prepares for plan/apply.
`inspect`	read	Shows detailed container or image information with structured state, image, and platform data.
`install`	write	Runs
`issue-close`	read	Closes an issue with an optional comment and reason. Returns structured data with issue number, state, URL, reason, and comment URL.
`issue-comment`	read	Adds, edits, or deletes a comment on an issue. Returns structured data with the comment URL, operation type, comment ID, issue number, and body echo.
`issue-create`	write	Creates a new issue. Returns structured data with issue number, URL, and labels applied.
`issue-list`	read	Lists issues with optional filters. Returns structured list with issue number, state, title, labels, assignees, author, creation date, and milestone.
`issue-update`	write	Updates issue metadata (title, body, labels, assignees, milestone, projects). Returns structured data with issue number and URL.
`issue-view`	read	Views an issue by number or URL. Returns structured data with state, labels, assignees, author, milestone, close reason, and body.
`jq`	read	Processes and transforms JSON using jq expressions. Accepts JSON from a file path or inline string.
`label-create`	write	Creates a new repository label. Returns structured data with label name, description, color, and URL.
`label-list`	read	Lists repository labels. Returns structured data with label name, description, color, and default status.
`lerna`	write	Runs Lerna monorepo commands (list, run, changed, version) and returns structured package information.
`lint`	read	Runs
`list`	read	Lists Go packages or modules and returns structured information.
`list-package`	read	Runs dotnet list package and returns structured NuGet package listings per project and framework.
`log`	write	Returns commit history as structured data.
`log-graph`	read	Returns visual branch topology as structured data. Wraps
`logs`	read	Retrieves container logs as structured line arrays.
`maven-build`	read	Runs
`maven-dependencies`	read	Shows the Maven dependency tree with structured output per artifact.
`maven-test`	read	Runs
`maven-verify`	read	Runs
`merge`	write	Merges a branch into the current branch. Supports abort, continue, and quit actions. Returns structured data with merge status, fast-forward detection, conflicts, and commit hash.
`mod-tidy`	destructive	Runs go mod tidy to add missing and remove unused module dependencies.
`mongosh-eval`	read	Evaluates a MongoDB expression via mongosh and returns the output.
`mongosh-stats`	read	Gets MongoDB database statistics (collections, objects, data size, storage, indexes) via mongosh.
`my-label`	read	A test label
`mypy`	read	Runs mypy and returns structured type-check diagnostics (file, line, severity, message, code).
`mysql-list-databases`	read	Lists all MySQL databases with structured output.
`mysql-query`	read	Executes a MySQL query and returns structured tabular output.
`network-ls`	read	Lists Docker networks with structured driver and scope information.
`nvm`	read	Manages Node.js versions via nvm.
`nx`	read	Runs Nx workspace commands and returns structured per-project task results with cache status.
`outdated`	read	Runs
`output`	read	Shows Terraform output values from the current state. Returns structured name/value/type/sensitive data.
`oxlint`	read	Runs Oxlint and returns structured diagnostics (file, line, column, rule, severity, message).
`package-clean`	read	Cleans Swift package build artifacts and returns structured result.
`package-init`	read	Initializes a new Swift package and returns structured result with created files.
`package-resolve`	read	Resolves Swift package dependencies and returns structured resolution results.
`package-show-dependencies`	read	Shows the dependency tree of a Swift package and returns structured dependency data.
`package-update`	write	Updates Swift package dependencies and returns structured update results.
`pip-audit`	read	Runs pip-audit and returns a structured vulnerability report.
`pip-install`	write	Runs pip install and returns a structured summary of installed packages.
`pip-list`	read	Runs pip list and returns a structured list of installed packages.
`pip-show`	read	Runs pip show and returns structured package metadata (name, version, summary, dependencies).
`plan`	read	Shows the Terraform execution plan with resource change counts. Read-only.
`playwright`	read	Runs Playwright tests with JSON reporter and returns structured results with pass/fail status, duration, and error messages.
`pm-ls`	read	Runs
`poetry`	read	Runs Poetry commands and returns structured output.
`post`	write	Makes an HTTP POST request via curl and returns structured response data.
`pr-checks`	read	Lists check/status results for a pull request. Returns structured data with check names, states, URLs, and summary counts (passed, failed, pending).
`pr-close`	read	Closes a pull request with an optional comment and optional branch deletion. Returns structured data with PR number, state, URL, and deleted-branch flag.
`pr-comment`	read	Adds, edits, or deletes a comment on a pull request. Returns structured data with the comment URL, operation type, comment ID, and body echo.
`pr-create`	write	Creates a new pull request. Returns structured data with PR number and URL.
`pr-diff`	write	Returns file-level diff statistics for a pull request. Use full=true for patch content.
`pr-list`	read	Lists pull requests with optional filters. Returns structured list with PR number, state, title, author, branch, labels, draft status, and merge readiness.
`pr-merge`	write	Merges a pull request by number, URL, or branch. Returns structured data with merge status, method, URL, and branch deletion status.
`pr-ready`	read	Marks a pull request as ready for review (or converts it back to draft when undo=true). Returns structured data with PR number, state, URL, and resulting draft status.
`pr-reopen`	read	Reopens a previously closed pull request, optionally with a comment. Returns structured data with PR number, state, and URL.
`pr-review`	read	Submits a review on a pull request (approve, request-changes, or comment). Returns structured data with the review event, URL, and body echo.
`pr-update`	write	Updates pull request metadata (title, body, labels, assignees, reviewers, milestone, base branch, projects). Returns structured data with PR number and URL.
`pr-view`	read	Views a pull request by number, URL, or branch. Returns structured data with state, checks, review decision, diff stats, author, labels, draft status, assignees, milestone, and timestamps.
`prettier-format`	write	Formats files with Prettier (--write) and returns a structured list of changed files.
`priority`	read	High priority issues
`ps`	read	Lists Docker containers with structured status, ports, and state information.
`psql-list-databases`	read	Lists all PostgreSQL databases via psql with owner, encoding, and size info.
`psql-query`	read	Executes a PostgreSQL query via psql and returns structured tabular output.
`publish`	write	Runs dotnet publish for deployment and returns structured output with output path and diagnostics.
`pull`	read	Pulls a Docker image from a registry and returns structured result with digest info.
`push`	write	Pushes commits to a remote repository. Returns structured data with success status, remote, branch, summary, and whether the remote branch was newly created.
`pyenv`	read	Manages Python versions via pyenv.
`pytest`	read	Runs pytest and returns structured test results (passed, failed, errors, skipped, failures).
`rebase`	read	Rebases the current branch onto a target branch. Supports abort, continue, skip, and quit for conflict resolution. Returns structured data with success status, branch info, conflicts, and rebased commit count.
`redis-command`	read	Executes a Redis command via redis-cli and returns the response.
`redis-info`	read	Gets Redis server info with structured sections (server, clients, memory, etc.).
`redis-ping`	read	Tests Redis connectivity by sending a PING command.
`reflog`	read	Returns reference log entries as structured data, useful for recovery operations. Also supports checking if a reflog exists.
`release-create`	write	Creates a new GitHub release with optional asset uploads. Returns structured data with tag, URL, draft, prerelease status, and assets uploaded count.
`release-list`	read	Lists GitHub releases for a repository. Returns structured list with tag, name, draft/prerelease/latest status, publish date, creation date, and URL.
`reload`	read	Rebuilds MCP server and sends tool list changed notification to refresh tool definitions.
`remote`	destructive	Manages remote repositories. Supports list (default), add, remove, rename, set-url, prune, and show actions. Returns structured remote data.
`remove`	destructive	Runs
`repo-clone`	read	Clones a GitHub repository. Returns structured data with success status, repo name, target directory, and message.
`repo-view`	read	Views repository details. Returns structured data with name, owner, description, stars, forks, languages, topics, license, and dates.
`request`	read	Makes an HTTP request via curl and returns structured response data (status, headers, body, timing).
`reset`	destructive	Resets the current HEAD to a specified state. Supports soft, mixed, hard, merge, and keep modes. The
`restore`	read	Runs dotnet restore to restore NuGet dependencies and returns structured results.
`rollup`	read	Runs Rollup bundler and returns structured bundle output with errors and warnings.
`rsync`	write	Syncs files between local and remote locations using rsync. Defaults to dry-run mode for safety.
`ruff-check`	read	Runs ruff check and returns structured lint diagnostics (file, line, code, message).
`ruff-format`	read	Runs ruff format and returns structured results (files changed, file list).
`run`	write	Runs a script or file with
`run-list`	write	Lists workflow runs with optional filters. Returns structured list with run ID, status, conclusion, and workflow details.
`run-rerun`	write	Re-runs a workflow run by ID. Optionally re-runs only failed jobs or a specific job. Returns structured result with run ID, status, and URL.
`run-view`	write	Views a workflow run by ID. Returns structured data with status, conclusion, jobs (with steps), and workflow details.
`search`	read	Searches the npm registry for packages matching a query.
`secret-delete`	destructive	Deletes a repository, organization, or environment GitHub Actions secret.
`secret-list`	read	Lists repository, organization, or environment secret names and metadata. Secret values are never exposed.
`secret-set`	write	Sets a repository, organization, or environment GitHub Actions secret. Secret values are sent via stdin and never returned.
`semgrep`	read	Runs Semgrep static analysis with structured rules and findings. Returns structured finding data with severity summary.
`shell`	read	Makes packages available in the environment and optionally runs a command.
`shellcheck`	read	Runs ShellCheck (shell script linter) and returns structured diagnostics (file, line, column, rule, severity, message).
`show`	write	Shows commit details and diff statistics for a given ref.
`ssh-keyscan`	read	Retrieves public host keys from a remote SSH server using
`ssh-run`	write	Executes a command on a remote host via SSH. Returns structured output with stdout, stderr, exit code, and duration.
`ssh-test`	read	Tests SSH connectivity to a remote host. Returns whether the host is reachable and any banner message.
`stash`	read	Pushes, pops, applies, drops, shows, or clears stash entries. Returns structured result with action, success, message, and stash reference.
`stash-list`	read	Lists all stash entries with index, message, date, branch, and optional file change summary. Returns structured stash data.
`state-list`	read	Lists all resources tracked in the Terraform state. Returns resource addresses.
`stats`	read	Returns a snapshot of container resource usage (CPU, memory, network/block I/O, PIDs) as structured data.
`status`	read	Returns the working tree status as structured data (branch, staged, modified, untracked, conflicts).
`stylelint`	read	Runs Stylelint and returns structured diagnostics (file, line, column, rule, severity, message).
`submodule`	write	Manages git submodules. Supports list (default), add, update, sync, and deinit actions. List returns structured submodule data with path, SHA, branch, and status.
`tag`	destructive	Manages git tags. Supports list (default), create, and delete actions. List returns structured tag data with name, date, and message. Create supports lightweight and annotated tags.
`task`	read	Runs a named task from deno.json via
`tasks`	read	Displays the tasks runnable from root project.
`test`	read	Runs
`tool-a`	read	Tool A
`tool-b`	read	Tool B
`tree`	read	Displays the dependency tree for a Rust project.
`trivy`	read	Runs Trivy vulnerability/misconfiguration scanner on container images, filesystems, or IaC configs. Returns structured vulnerability data with severity summary.
`tsc`	read	Runs the TypeScript compiler and returns structured diagnostics (file, line, column, code, message).
`turbo`	read	Runs Turborepo tasks and returns structured per-package results with cache hit/miss info.

Trust audit

SAFEgrade C · trust 73/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

Layer	What it checks	Result
L0	Provenance & inventory	PASS
L1	Static analysis of the code	PASS
L2	Instruction surface (what it tells the agent)	PASS
L3	Class-specific surface	WARN
L4	Behavioural (sandbox)	SKIPPED

What the source does

Filesystem: declared (1 observation(s))
Network: declared (3 observation(s))
Shell: declared (4 observation(s))
Dependencies: not all pinned
Secrets in source: none-found

Findings (25)

MEDIUMFilesystem / path · mcp.destructive_tools · CWE-22, CWE-59

mod-tidy, remote, remove, reset, secret-delete, tag, vagrant, variable-delete, workspace

Why it matters. 9 tool(s) can delete or overwrite

Fix. prefer a read-only mode or scoped tokens; the page states the blast radius

LOWInventory / provenance · inv.binary · CWE-1104

tests/smoke/fixtures/git/log/s01-default.txt

s01-default.txt