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 commands/lugassawan/swe-workbench/report-issuegit clone --depth 1 https://github.com/lugassawan/swe-workbenchWrote 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/commands/lugassawan/swe-workbench/report-issue)<a href="https://agentmods.dev/commands/lugassawan/swe-workbench/report-issue"><img src="https://agentmods.dev/badge/commands/lugassawan/swe-workbench/report-issue.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 | $0.00019 | $0.04694 |
| Opus 5 | $0.00010 | $0.02347 |
| Sonnet 5 | $0.00004 | $0.00939 |
| Haiku 4.5 | $0.00002 | $0.00469 |
Grade B, and why
report-issue scanned grade B with 1 finding 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 3d 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
- **Claude Code — plugin version:** list `~/.claude/plugins/cache/swe-workbench/swe-workbench/` version directories, sort semantically (`sort -V`), take the highest with `tail -1`, then read `plugin.json` from it: `ls ~/ How it starts
The opening of the file, as written. The whole thing — 176 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Override (this command only): The
swe-workbench:product-manageragent's default rules "Does not pass--repoexplicitly togh issue create" (Decision boundaries) and "Never pass--repo" (Mutation rules) are suspended for/swe-workbench:report-issue. You MUST pass--repo lugassawan/swe-workbenchon everyghinvocation in this flow. No other rule is changed.
The user wants to file a plugin issue: $ARGUMENTS
Target repo (hardcoded): repo="lugassawan/swe-workbench"
Step 0 — Blank-argument routing
If $ARGUMENTS is empty, do NOT error. Instead:
Step 0a — Mode selector
Ask the user to choose a mode, then wait for the reply:
How would you like to find a plugin-related thought to report?
1. Quick pick — up to 3 raw candidates from this session/memory
2. Synthesize — aggregate all memory into ranked, themed enhancement insights
Reply with 1 (quick pick) or 2 (synthesize).
Reply 1 → Branch A — Quick pick. Reply 2 → Branch B — Synthesize.
Branch A — Quick pick
-
Scan the current conversation for actionable plugin-related thoughts: frustrations, lessons learned, complaints about plugin behaviour, feature ideas, or pain points. Collect up to 3 candidates with a one-line framing each.
-
If conversation yields fewer than 3, scan memory at
~/.claude/projects/<project-slug>/memory/MEMORY.md(where<project-slug>is derived from the current working directory path by replacing each/with-, then stripping any resulting leading-; e.g./Users/foo/bar→Users-foo-bar). Follow symlinks tofeedback_*.mdandproject_*.mdentries referenced there. Collect additional candidates until you have up to 3, prioritising entries that mention the plugin, commands, agents, or skills by name. -
Present candidates numbered 1–N (max 3) with a one-line framing each:
I found a few plugin-related thoughts from this session/memory. Pick one or type your own: 1. <one-line framing> 2. <one-line framing> 3. <one-line framing> Reply with 1/2/3 or type your own thought.Wait for the user's reply. The selected or typed thought becomes
$ARGUMENTSand flow continues at the Auth + repo check step below.
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.
- 3d ago First seen · 176 lines · 19 tokens per session scan B 6f8dcd51bf0a
report-issue is a command published in the GitHub repository lugassawan/swe-workbench (2 stars, last pushed 3d ago), licensed MIT. It adds 19 tokens to every session and 4,694 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
brooks-audit
Run a Brooks-Lint architecture audit.
brooks-health
Run a Brooks-Lint codebase health dashboard across all four dimensions.
add-technology
Ajouter une nouvelle technologie à claude-craft avec les best practices de Context7 et recherche web.
daily-standup
Génération Résumé Daily Stand-up.
generate-component
Generate Angular standalone component with tests.
init
Bootstrap de la structure Claude-Craft pour un nouveau projet.