depwire-cli - npm Package Compare versions

Comparing version 1.8.6 to 1.8.7

package.json

 {
   "name": "depwire-cli",
-  "version": "1.8.6",
+  "version": "1.8.7",
   "description": "Dependency graph + 23 MCP tools for AI coding assistants. Impact analysis, health scoring, security scanner, agent coordination.",
@@ -5,0 +5,0 @@ "type": "module",

README.md

@@ -72,2 +72,26 @@ # Depwire
 
+## Benchmark Results
+
+Tested on [payloadcms/payload](https://github.com/payloadcms/payload) — 645 files, 9,292 symbols, 80-file cascading refactor (adding a required parameter to a core error class used across the entire codebase).
+
+| Mode | Duration | API Calls | Tokens | Cost | Correctness |
+|------|----------|-----------|--------|------|-------------|
+| Without Depwire | 16m 46s | 40 | 2,959,169 | $9.03 | 100% |
+| Depwire (no guidance) | 11m 20s | 46 | 3,399,822 | $9.80 | 100% |
+| Depwire + workflow | **10m 43s** | **26** | **2,165,356** | **$7.35** | 100% |
+
+**Depwire + guided workflow vs no Depwire:**
+- 36% faster
+- 35% fewer API calls
+- 27% fewer tokens
+- 19% lower cost
+
+> **Key finding:** Depwire without an explicit workflow prompt performed worse than no Depwire — the agent had tools available but didn't use them. The improvement only manifests when agents follow the `depwire prompt` workflow. Run `depwire prompt` to get the proven workflow for your agent.
+
+> **How it works:** `affected_files src/errors/APIError.ts` returned the complete 84-file blast radius in one tool call, eliminating all exploratory loops. The agent went straight to fixing files instead of discovering them.
+
+[Full benchmark methodology and scripts →](https://github.com/depwire/depwire-benchmark)
+
+---
+
 ## The problem
@@ -132,2 +156,3 @@
 | [rstudio/plumber](https://github.com/rstudio/plumber) | R | 197 | 1,194 | 219 | — |
+| [payloadcms/payload](https://github.com/payloadcms/payload) | TypeScript | 645 | 9,292 | 3,511 | — |
 
@@ -330,2 +355,3 @@ > Numbers from real `depwire parse` runs on public repositories. Last validated: v1.8.2 (June 2026).
 | `depwire parse` | Parse and export dependency graph as JSON |
+| `depwire prompt` | Get the proven workflow prompt for your AI agent |
 | `depwire diff` | Structural diff between two git commits — symbols, edges, health, security |
@@ -336,2 +362,14 @@ | `depwire mcp` | Start MCP server for AI coding assistants |
 
+### `depwire prompt` — proven workflow for AI agents
+
+```bash
+# Get the proven workflow prompt for your agent
+depwire prompt                    # generic
+depwire prompt --tool claude      # Claude Code optimized
+depwire prompt --tool cline       # Cline optimized
+depwire prompt --tool codex       # Codex optimized
+```
+
+The guided workflow reduced cost by 19% and time by 36% in benchmarks. Paste the output as your agent's system context before starting any complex task.
+
 ---
@@ -374,2 +412,16 @@
 
+### Auto-generated project context
+
+After running `depwire parse .`, Depwire generates `.depwire/AGENTS.md` — a project-specific context file containing module structure, key files, health summary, and MCP quick-start commands.
+
+Claude Code reads `AGENTS.md` automatically when present. Add it to your `CLAUDE.md`:
+
+```bash
+# In your project root CLAUDE.md:
+echo "## Depwire Context" >> CLAUDE.md
+echo "Read .depwire/AGENTS.md for codebase architecture." >> CLAUDE.md
+```
+
+This gives every Claude Code session instant orientation without any tool calls — the benchmark showed this alone reduces exploratory iterations.
+
 ![Claude Desktop with Depwire MCP](./assets/claude.gif)
@@ -639,2 +691,4 @@
 - Fast MCP startup — loads from depwire-output.json in <100ms regardless of project size
+- `depwire prompt` — proven workflow prompt for AI agents (19% cheaper, 36% faster in benchmarks)
+- Auto-generated `.depwire/AGENTS.md` project context after `depwire parse`
 
@@ -641,0 +695,0 @@ **Coming next**

New alerts

Long strings

Supply chain risk

Contains long string literals, which may be a sign of obfuscated or packed code.

Unidentified License

License

Something that seems like a license was found, but its contents could not be matched with a known license.

Fixed alerts

Major refactor

Supply chain risk

Package has recently undergone a major refactor. It may be unstable or indicate significant internal changes. Use caution when updating to versions that include significant changes.

Long strings

Supply chain risk

Contains long string literals, which may be a sign of obfuscated or packed code.

Unidentified License

License

Something that seems like a license was found, but its contents could not be matched with a known license.

Improved metrics

Total package byte prevSize

26717633

0.02%

Lines of code

27411

0.15%

Number of lines in readme file

744

7.83%

Number of medium supply chain risk alerts

9

-10%

Worsened metrics

Number of low supply chain risk alerts

66

3.13%

No dependency changes