AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Code Intelligence

skill-dachent-skills-code-intelligence · by dachent

use only in Claude Code as an explicit provider router when Graphify, code-mapper, or selective CodeQL availability must be assessed. do not install or auto-trigger for Codex GPT-5.6 Sol; native direct inspection, Plan Mode, explorer work, and subagents cover this routing without an added analyzer.

— No reviews yet
0 installs
0 views
— view→install

Install

$ agentstack add skill-dachent-skills-code-intelligence

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • ✓ Prompt-injection patterns
  • ✓ Secret / credential exfiltration
  • ✓ Dangerous shell & filesystem operations
  • ✓ Untrusted network calls
  • ✓ Known-malicious package signatures

What it can access

  • ✓ Network access No
  • ✓ Filesystem access No
  • ● Shell / process execution Used
  • ✓ Environment & secrets No
  • ✓ Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-dachent-skills-code-intelligence)

Reliability & compatibility

✓ Security review passed
0 installs to date
— no reviews yet
● 26d ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Code Intelligence? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Code Intelligence — Claude Code Provider Router

Purpose

Use only as a Claude Code provider router. Choose the least expensive installed provider that can answer the question reliably, preserve evidence provenance, and verify material conclusions against current source and tests.

Do not install or auto-trigger this skill for Codex GPT-5.6 Sol. Native direct inspection, Plan Mode, explorer work, and subagents already cover the routing function without adding an analyzer.

Keep providers independent. This skill does not vendor or absorb Graphify, code-mapper-skill, or CodeQL.

Active Routes

  1. Direct source and tests — default for known non-Python targets, small repositories, stale or missing graph data, and final verification.
  2. Graphify — broad or unknown-target discovery only when an existing graph is relevant and fresh.
  3. code-mapper-skill — Python imports, callers, cycles, blast radius, artifacts, contracts, APIs, catalogs, and lineage.
  4. Selective CodeQL through code-mapper-skill — Python security or value-flow questions when CodeQL is available and the mapper's documented trigger applies.

Codebase-Memory is a benchmark candidate, not an active provider. Serena, Joern/codebadger, SCIP/Sourcegraph, Semgrep, and Aider-style maps are not part of the MVP routing surface.

Workflow

  1. Confirm this is an explicit Claude Code provider-routing request, not ordinary repository inspection or planning.
  2. Classify the request as broad discovery, known-target Python analysis, artifact or contract lineage, security-flow triage, or direct source work.
  3. Run scripts/preflight.py only when Graphify freshness or local capability matters. It must not parse graph.json merely to choose a route.
  4. Apply references/routing-policy.md. Prefer the known-target fast path and do not build an index unless repeated discovery justifies it.
  5. Follow references/freshness-policy.md before graph-backed work. Graph existence is not proof of freshness.
  6. Follow references/provider-contract.md and record route, freshness, latency, evidence, warnings, and fallback use.
  7. Verify against current source and tests.
  8. Report stale evidence, unavailable capabilities, inferences, and unverified assumptions.

Fast Rules

  • Known Python file or symbol: code-mapper-skill, without Graphify I/O.
  • Python artifact, contract, API, catalog, or lineage question: code-mapper-skill.
  • Python security/value-flow question: code-mapper-skill with selective CodeQL when available; otherwise report the limitation.
  • Broad unknown-target question: fresh Graphify if already available; otherwise direct source exploration.
  • Known non-Python target or small repository: direct source and host-native tools.
  • Durable planning is outside this router; use the harness-native planning workflow.

Performance Boundaries

The direct route to code-mapper-skill must not read or parse Graphify's graph and must not spawn Graphify. The router itself adds no runtime overhead: routing is a decision tree applied in-context, not a process. route.py is an importable policy function (decide_route), never a subprocess in the request path — there is no CLI to spawn between the request and the selected provider.

Required References

  • references/architecture.md
  • references/routing-policy.md
  • references/freshness-policy.md
  • references/provider-contract.md
  • references/alternatives.md
  • references/implementation-plan.md
  • references/adversarial-review.md
  • ADR files

Bundled Scripts

  • scripts/route.py — importable routing policy (decide_route); no CLI, not spawned at runtime
  • scripts/preflight.py — Graphify freshness / capability check, run only when graph-backed work is in scope
  • scripts/test_route.py
  • scripts/test_preflight.py

Run scripts with Python 3.10 or newer. Do not report repository work complete until the PR's required checks have concluded successfully.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.