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/lugassawan/swe-workbench/lspgit clone --depth 1 https://github.com/lugassawan/swe-workbenchWhat 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 | $0.00000 | $0.00473 |
| Opus 5 | $0.00000 | $0.00236 |
| Sonnet 5 | $0.00000 | $0.00095 |
| Haiku 4.5 | $0.00000 | $0.00047 |
Grade A, and why
lsp 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 2d 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.
What it actually says
LSP navigation
bin/swe-workbench-lsp gives you a real language server's semantic index of
the codebase — the same engine behind an IDE's "Go to Definition" or "Find
All References," resolving symbols by type and scope rather than by
spelling. Reachable from Bash on any harness, since it never depends on a
harness-provided LSP tool being wired up for subagents. It exposes eight
navigation subcommands — refs, def, impl, callers, callees, hover,
symbols, wsymbols — plus check for availability (see below).
It follows; it does not find
The script has no free-text search of its own — every call needs an anchor position first. The handoff is a fixed two-step pair:
- Search the codebase (
Grep/Glob, or any equivalent text search) to locate the anchor — the symbol's declaration or a call site — giving you its file path and, ideally, its exact name. - Feed that anchor to the script:
swe-workbench-lsp def <file>:<line>orswe-workbench-lsp refs <file> --symbol <name>to expand outward from it, orcallers/calleesto walk the call graph.
Text search is weakest exactly where this matters: shadowed names, same-named methods on unrelated types, re-exports, and callers reached only through an interface all read as text noise to a grep but resolve correctly through the language server's semantic index.
Availability gate — mandatory
Run
swe-workbench-lsp checkonce at the start of a task that will need symbol navigation — it only confirms the server binary is onPATH, not that a real handshake with your project succeeds. If the extension you need isn'tOK(exit 3 from any subcommand, orMISSING/absent fromcheck's output), stateLSP unavailable — falling back to Greponce and use Grep for the remainder of this run. Do not retry.
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.
- 2d ago First seen · 36 lines · 0 tokens per session scan A fa3a63204e1a
lsp is an agent published in the GitHub repository lugassawan/swe-workbench (2 stars, last pushed 2d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 473 tokens. 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 2026-08-31.
Other agents, from other repositories
architect-review
Master software architect specializing in modern architecture patterns, clean architecture, microservices, event-driven systems, and DDD. Reviews system designs and code changes for architectural integrity, scalability, and maintainability. Use PROACTIVELY for architectural decisions.
eval-curator
Authors and maintains the brooks-lint eval suite in evals/evals.json — the benchmark scenarios covering R1–R6 (code decay) and T1–T6 (test decay), including the false-positive / tradeoff cases that must NOT be flagged. Ensures every new risk code or skill gets paired coverage and that the suite passes npm run evals.…
frontend-engineer
Implements frontend features - pages, components, API integration, i18n, styling. Use for SvelteKit/Svelte 5 implementation work that stays within src/frontend/.
ux-designer
For website changes, owns the "spec" phase as a design spec — information architecture, wireframe/description of the change, interaction notes. Gate 1 approves design, not prose. Persona frontend-ux.
php-reviewer
PHP 8.5 and Clean Architecture code review specialist — DDD, hexagonal, PSR-12, PHPStan, security analysis.
react-reviewer
React 19.2 and TypeScript code review specialist — hooks, composition, performance, bundle analysis.