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 instructions/ghostwright/specter/claude-mdgit clone --depth 1 https://github.com/ghostwright/specterWhat 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 | $0.01884 | $0.01884 |
| Opus 5 | $0.00942 | $0.00942 |
| Sonnet 5 | $0.00377 | $0.00377 |
| Haiku 4.5 | $0.00188 | $0.00188 |
Grade C, and why
specter CLAUDE.md scanned grade C with 2 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
make clean # rm -rf bin/ dist/ Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl https://test.yourdomain.com/health How it starts
The opening of the file, as written. The whole thing — 171 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Specter
Go CLI that deploys AI agent VMs on Hetzner Cloud in 90 seconds. Interactive TUI (Bubbletea v2). Automatic DNS (Cloudflare) and TLS (Let's Encrypt via Caddy). Apache 2.0.
Build
make build # -> bin/specter
make lint # go vet + gofmt check
make test # go test ./...
make ci # lint + build (same as CI workflow)
make install # go install to $GOPATH/bin
make clean # rm -rf bin/ dist/
Project Structure
cmd/specter/
main.go Entry point, Cobra root
commands/
root.go Root command, global flags
deploy.go VM creation + DNS + SSH provisioning + TLS
destroy.go VM + DNS teardown
init.go Setup wizard (tokens, firewall, server cache)
image.go Golden snapshot build + list
list.go Agent inventory with health checks
status.go Single agent status
ssh.go SSH into VM
logs.go systemd journal streaming
update.go Agent restart
version.go Version output
internal/
cloudflare/client.go Cloudflare DNS API (raw HTTP, not SDK)
config/
config.go ~/.specter/config.yaml management
server_types.go Hetzner server type cache + fuzzy match
state.go Agent state persistence
hetzner/client.go Hetzner Cloud API (hcloud-go v2)
templates/
cloudinit.go Cloud-init user-data template [FROZEN]
systemd.go systemd unit template [FROZEN]
caddyfile.go Caddy reverse proxy template [FROZEN]
tui/
app.go Main Bubbletea model (~1,300 lines)
agent_list.go Dashboard list view
agent_detail.go Agent detail panel
deploy_form.go Deploy form (huh v2)
deploy_model.go Deploy data types
deploy_progress.go Deploy phase progress
image_build.go Image build progress
setup_wizard.go First-run setup
logs_viewport.go Log viewer
confirm_dialog.go Confirmation dialogs
help_overlay.go Keyboard help
status_bar.go Bottom bar
dashboard_styles.go Lipgloss styles
messages.go Bubbletea messages
theme.go Color palette
pkg/version/version.go Version vars (set via ldflags)
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 · 171 lines · 1,884 tokens per session scan C 8293a2ca5159
specter CLAUDE.md is an instructions file published in the GitHub repository ghostwright/specter (39 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 1,884 tokens to every session, about $0.0094 per session on Opus 5. A static security scan graded it C with 2 findings (recursive force delete, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
OpenSail CLAUDE.md
Instructions for TesslateAI/OpenSail, covering critical rule: investigation is read-only, research folder, commit messages, banned - describes development steps, not the diff and good - describes what the full diff actually contains.
ecspresso CLAUDE.md
Instructions for kayac/ecspresso, covering claude.md, project overview, git workflow, build commands and build the binary.
mcp-ssh-manager CLAUDE.md
Instructions for bvisible/mcp-ssh-manager, covering claude.md, project overview, architecture, commands and setup and installation.
moosestack copilot-instructions.md
Instructions for 514-labs/moosestack, covering moose repository guide, code standards, required before each commit, development flow and repository structure.
agentscope-runtime copilot-instructions.md
Instructions for agentscope-ai/agentscope-runtime, covering agentscope runtime code review guide, 1. code quality, 2. [must] code security, 3. [must] testing & dependencies and 4. code standards.
do-app-platform-skills CLAUDE.md
Instructions for digitalocean-labs/do-app-platform-skills, covering claude.md, project overview, repository structure, architecture and skill routing.