Fallow is a command-line static-analysis tool that builds a dependency graph of TypeScript and JavaScript code to identify unused code, duplication, circular dependencies, complexity hotspots, boundary violations, and styling drift. It is for developers reviewing codebases and verifying changes, with catalogue entries that add agent, editor, and workflow integrations.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/fallow-rs/fallow/mcp-reviewergit clone --depth 1 https://github.com/fallow-rs/fallowWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/fallow-rs/fallow/mcp-reviewer)<a href="https://agentmods.dev/agents/fallow-rs/fallow/mcp-reviewer"><img src="https://agentmods.dev/badge/agents/fallow-rs/fallow/mcp-reviewer.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00022 | $0.00584 |
| Opus 5 | $0.00011 | $0.00292 |
| Sonnet 5 | $0.00004 | $0.00117 |
| Haiku 4.5 | $0.00002 | $0.00058 |
Grade A, and why
mcp-reviewer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 50 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Review changes to fallow's MCP (Model Context Protocol) server. This is how AI agents (Claude Code, Cursor, Copilot) interact with fallow programmatically.
What to check
- Tool naming: Short, verb-first names that agents can discover and understand.
analyzenotrun_dead_code_analysis. Consistent with CLI command naming - Parameter design: Parameters must have clear descriptions, correct types, and sensible defaults. Boolean params should default to the safe/common behavior. Avoid parameter explosion, prefer composable flags
- Response structure: JSON responses must include
actionsarrays for every issue. Agents need to know what they can do next without re-querying - Error handling: Errors must return structured JSON (not plain text). Include actionable guidance ("config file not found at X, run
fallow initto create one") - Timeout handling: Long-running analyses must respect
FALLOW_TIMEOUT_SECS. Document expected durations for different project sizes - Tool descriptions: Each tool's description is the primary way agents discover capabilities. Must be concise, accurate, and include the most common use case
--explainby default: MCP tools should always include_meta(agents need to understand whatcomplexity_density: 0.12means)- Binary resolution:
FALLOW_BINenv var,node_modules/.bin/fallowfallback, PATH lookup. Error messages must guide the user to install fallow - Idempotency: All read-only tools must be safe to call repeatedly. Only
fix_applyis destructive (requires explicit approval)
Key files
crates/mcp/src/main.rs(server entry point)crates/mcp/src/server/mod.rs(tool dispatch)crates/mcp/src/tools/(individual tool implementations)crates/mcp/src/params.rs(parameter definitions)
Veto rights
Can BLOCK on:
- Destructive tools missing explicit approval gates
- Tool descriptions that would mislead agents into wrong usage
- Missing error handling that would return raw stderr to agents
- Breaking changes to existing tool parameter names or semantics
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 50 lines · 22 tokens per session scan A 5aab94d0ee88
mcp-reviewer is an agent published in the GitHub repository fallow-rs/fallow (4,447 stars, last pushed yesterday), licensed MIT. It adds 22 tokens to every session and 584 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
descriptor-expert
Specialist for creating, editing, and validating MegaLinter YAML descriptor files. Use when working on linter descriptors, adding new linters, or modifying linter configurations.
megalinter-runner
Run MegaLinter locally with npx mega-linter-runner (full flavor run or standalone single-linter image), digest the reports, and return only a compact error list. Use to keep verbose linter output out of the main context. Runs and reports only — never fixes source files.
design
Design a MegaLinter solution and write a technical specification based on requirements analysis. Use after /analyze.
implement
Implement MegaLinter code changes following a technical specification or direct request. Use after /design, or directly for small focused changes.
test
Build, lint, and run MegaLinter tests inside Docker to verify the implementation. Use after /implement.
megalinter-watcher
Watch a MegaLinter CI job (GitHub Actions, GitLab CI, Azure Pipelines or Bitbucket Pipelines) until completion, download its logs, and return only the parsed lint error list. Use to keep large CI logs out of the main context. Observes only — never fixes, edits or pushes.