pypeeker-cli 1.6.0

A zero-dependency Python CLI toolset designed for AI agents to analyze codebases with surgical precision.

pypeeker-cli

Agentic DX for Python. DX = developer experience: the tools, feedback, and ergonomics that make working in a codebase feel good. Agentic DX is the same idea, but for the AI agent doing the work instead of a human. pypeeker gives AI coding agents AST-level views of any Python codebase — API surfaces, side-effect maps, import cycles, symbol locations — surfaced through MCP in token-efficient structured answers. So your agent can navigate, reason about, and refactor Python without reading whole files.

Read-only by design. pypeeker never edits, renames, or rewrites your code. No apply, no fix, no auto-refactor surface. Every tool answers a question about the codebase; the agent acts on the answer. This is a hard boundary, not a roadmap gap.

---

Cartograph Showcase

pypeeker-cli is a premier showcase for Cartograph, a platform for reusable engineering.

Every core feature in pypeeker-cli, from AST parsing to graph cycle detection, is implemented as a standalone, validated Cartograph widget. This architecture ensures that pypeeker-cli is not just a tool, but a modular assembly of hardened building blocks that can be easily extended or repurposed.

---

What the agent gets

Three tools over MCP, designed to keep the agent's context budget tight:

audit(directory, kind=...) — project-wide checks

• kind="cycles" — import cycles + a cycle_hubs ranking of the most-tangled files
• kind="missing-imports" — hallucinated or broken internal import paths
• kind="interfaces" — missing docstrings and type annotations

peek(path, mode=..., symbol=...) — file or symbol inspection

• mode="skeleton" — file/package API surface with line ranges for targeted reads
• mode="locate" / "usages" / "ancestry" — find a symbol's definition, usages, or class parents
• mode="impact" — function side-effect map; set depth=N + root for transitive blast-radius across files

cli(command, args) — escape hatch

Runs pypeeker <command> <args> and returns stdout. Use when you need a flag the consolidated tools don't expose (e.g., --include-deps, custom --ignore, --format json on a tool that defaults to text).

---

Tool Showcase

Two views the Read tool can't give an agent in one call — both surfaced through peek:

• mode="skeleton" — the API surface, without function bodies. Map a file or a whole package, get every signature with line ranges, jump straight to the methods that matter.
• mode="impact" — the side-effect map of a function, without mentally tracing the body. What it calls, what it reads, what it writes, whether it touches globals — answered structurally.

The walkthrough below uses psf/requests — a real, well-known library — to answer a real agent question: "How does requests handle authentication during redirects?"

Map the file — peek with mode="skeleton"

The agent calls peek over MCP with skeleton mode. The tool strips every function body and returns imports, signatures, docstrings, and the line range each symbol occupies.

Agent invokes: peek(path="requests/sessions.py", mode="skeleton") Returns:

import os
import time
from .auth import _basic_auth_str
from .cookies import RequestsCookieJar, extract_cookies_to_jar
# ...

class SessionRedirectMixin:  # L107-353
    def get_redirect_target(self, resp):  # L108-126
        """Receives a Response. Returns a redirect URI or ``None``"""
        ...

    def should_strip_auth(self, old_url, new_url):  # L128-158
        """Decide whether Authorization header should be removed when redirecting"""
        ...

    def rebuild_auth(self, prepared_request, response):  # L282-300
        """When being redirected we may want to strip authentication..."""
        ...

The two methods we want — should_strip_auth and rebuild_auth — are now visible with their exact line ranges.

Trace what a function touches — peek with mode="impact"

Once the agent has found the function, the next question is usually "what does this affect?" — globals, class state, external calls. Reading the body to find out is the manual approach. peek in impact mode answers structurally:

Agent invokes: peek(path="requests/sessions.py", mode="impact", symbol="SessionRedirectMixin.rebuild_auth") Returns:

{
  "external": {
    "calls":  ["get_netrc_auth", "prepared_request.prepare_auth", "self.should_strip_auth"],
    "reads":  ["self.trust_env", "..."],
    "writes": [],
    "globals": []
  },
  "internal": {
    "writes": ["headers", "new_auth", "url"]
  }
}

Three external dependencies. Zero global writes. Zero shared-state mutation. That's a refactor-safety check the agent can do before changing anything, in one call instead of a multi-pass read.

Impact mode accepts both bare names (rebuild_auth) and qualified names (SessionRedirectMixin.rebuild_auth) for unambiguous targeting when multiple classes share method names.

depth=N: transitive blast radius across files

The harder refactor question is "if I change this method's contract, what's the cascade?" — that's what its callees touch, and what their callees touch, and so on. depth=N walks the call graph up to N levels (max 5), aggregates every external write, global mutation, and reached symbol into one answer:

Agent invokes: peek(path="backend/services/chat_service.py", mode="impact", symbol="ChatService.process_message", depth=2, root=".") Returns:

# impact: ChatService.process_message  (depth 2)

transitive surface across 32 reached symbols:
  external calls:   323 unique
  external writes:  1       <- danger zone for refactors
    self._failed_edit_hashes  in ChatService.process_message (depth 0)
  globals modified: 0

reached symbols:
  depth 0  ChatService.process_message     backend/services/chat_service.py
  depth 1  build_summary_context            backend/context_builder.py
  depth 1  schedule_summary_update          backend/summary_updater.py
  ... (29 more)

unresolved (notable, 63):
  print                          in ChatService.process_message [effect_builtin]
  asyncio.create_task            in ChatService.process_message [stdlib]
  re.search                      in ChatService.process_message [stdlib]
  hashlib.md5                    in ChatService.process_message [stdlib]
  json.dumps                     in ChatService.process_message [stdlib]
  ...

unresolved (filtered, 316): 265 dispatch, 48 pure builtin, 3 exception

A 2,805-line method that calls into 32 other functions transitively mutates exactly one external attribute. That's a structural answer to "is this safe to refactor?" — derived in one tool call, with the unresolved cases flagged so the agent knows where its view is incomplete.

Static resolution only. self.X, Class.X, and statically-imported names are followed. Dynamic dispatch (obj.method() where obj's type is inferred at runtime) is reported as unresolved, not silently followed. Cycle-safe (visited-set deduplication). Project-bounded via the root parameter.

The token math

Same task — find where requests handles redirect auth — measured against actual agent tool output, including the Read tool's line-number prefixes. Tokens counted with OpenAI's o200k_base tokenizer (used by GPT-4o, o1, o3, and GPT-5).

Workflow	Tokens
Read sessions.py (full file)	9,063
skeleton sessions.py + 2 targeted Read calls (L128-158, L282-300)	3,218

The agent ends up reading only the two methods that actually matter, instead of 833 lines hoping the relevant code is in there — and the skeleton's line ranges (# L128-158) made the targeting possible.

For project-scale orientation, mapping the full requests package (18 files):

Operation	Tokens
Read every .py file in requests/	59,901
skeleton requests/	19,508

The entire library API surface — every public class, signature, docstring, and line range — in one call.

Also covered by the same tools

• peek(mode="locate"|"usages"|"ancestry") — AST-aware symbol search with scope ranges. No false positives from substring matches; distinguishes definitions from usages from inheritance.
• audit(kind="cycles") — import cycles, separates runtime cycles from safe TYPE_CHECKING cycles, ranks files by cycle membership.
• audit(kind="missing-imports") — hallucinated or broken internal imports.
• audit(kind="interfaces") — missing docstrings and type annotations across a project.

All modes default to condensed text/stub output for the agent. Pass format="json" when the agent needs to navigate fields programmatically.

---

Install

Two steps: install the package, then register it with your agent. The package ships an MCP server (pypeeker mcp); your agent connects to that server and gets the tools described above.

pip install pypeeker-cli

After install, register with whichever agent you use:

Gemini CLI

Install as a native extension:

gemini extensions install https://github.com/benteigland11/pypeeker

Claude Code

Add as a persistent MCP server:

claude mcp add pypeeker-cli -- pypeeker mcp

Codex

Add as a global MCP server:

codex mcp add pypeeker-cli -- pypeeker mcp

Cursor and Roo Code

Pypeeker is pre-configured for automatic detection via .cursor/mcp.json.

Continue

The server is pre-configured via .continue/mcpServers/pypeeker.json.

Windsurf

The server is pre-configured via .windsurf/mcp_config.json.

Claude Desktop and Aider

Copy the JSON from claude_desktop_snippet.json into your global claude_desktop_config.json.

Update notifications

When the MCP server starts, it checks PyPI once per session (cached 24h, 2-second timeout, fail-quiet) for a newer release. If your installed version is stale, agents see a one-line pip install -U pypeeker-cli notice prepended to the server instructions; up-to-date sessions see nothing. Set PYPEEKER_CLI_NO_VERSION_CHECK=1 to disable the check entirely (e.g. on airgapped networks).

---

Reusable Core

Built using the following Cartograph Widgets (found in cg/):

• cg-infra-agent-cli-python: Machine-first declarative CLI framework.
• data-ast-import-parser-python: Root-aware static import analysis.
• data-ast-skeleton-parser-python: API signature extraction.
• data-ast-symbol-locator-python: Surgical symbol pinpointing and ancestry.
• data-ast-interface-validator-python: API gap detection.
• data-ast-impact-analyzer-python: Side-effect and dependency mapping.
• infra-mcp-manifest-generator-python: Automated distribution scaffolding.
• universal-agent-response-python: Standardized JSON schema.
• universal-list-paginator-python: Result set pagination.

---

License

Licensed under the Apache License, Version 2.0. See the LICENSE file for details.