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/bitsky-tech/amphiloop/buildgit clone --depth 1 https://github.com/bitsky-tech/AmphiLoopWhat 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 | $0.00140 | $0.01715 |
| Opus 5 | $0.00070 | $0.00857 |
| Sonnet 5 | $0.00028 | $0.00343 |
| Haiku 4.5 | $0.00014 | $0.00171 |
Grade A, and why
build 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 yesterday.
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 — 120 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Build Pipeline
Turn any task into a working bridgic-amphibious project. The pipeline is domain-agnostic by default, with optional pre-distilled domain context injected per supported domain (browser, ...).
Argument parsing
$ARGUMENTS may be empty, or contain a single domain flag in either form --<domain> (e.g., --browser). Trim whitespace; ignore case.
- Flag present → set
SELECTED_DOMAIN = <domain>and skip auto-detection. Validate that{PLUGIN_ROOT}/domain-context/<domain>/exists. If it does not, list the available domains (subdirectories of{PLUGIN_ROOT}/domain-context/) and ask the user to pick one or rerun without a flag. - No flag → leave
SELECTED_DOMAINunresolved; resolve it during Phase 1's auto-detection step.
Anything else in $ARGUMENTS (extra tokens, multiple flags) → stop and ask the user to clarify.
Pipeline Workflow
1. Initialize Task (this command — generate TASK.md template, user fills in; then auto-detect domain if not flagged)
2. Configure & Setup (this command — inline methodology; user interactions; writes build_context.md)
3. Exploration (this command — inline methodology; may invoke human-in-the-loop)
4. Generate Amphibious Code (→ amphibious-code subagent)
5. Verify (this command — inline methodology; may invoke human-in-the-loop)
Path variables:
{PLUGIN_ROOT}and{PROJECT_ROOT}are path placeholders — all paths below use these prefixes. If either is missing, the plugin was not loaded correctly — do not proceed. build_context_path — always{PROJECT_ROOT}/.bridgic/build_context.md. domain_context_path —{PLUGIN_ROOT}/domain-context/<SELECTED_DOMAIN>/<phase>.mdwhenSELECTED_DOMAINis resolved, otherwise the literalnone(generic flow).<phase>isexplore.mdfor Phase 3,code.mdfor Phase 4,verify.mdfor Phase 5.
After Phases 3 and 4, refresh build_context.md in two places:
- Outputs — replace the matching
(filled by Phase N)placeholder with the phase's primary output path. - env_ready — read
{PROJECT_ROOT}/pyproject.tomland update the dump under--- pyproject.toml ---inside theenv_ready:block with its current contents.
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.
- yesterday First seen · 120 lines · 140 tokens per session scan A 8755e0b8407a
build is a command published in the GitHub repository bitsky-tech/AmphiLoop (68 stars, last pushed 3mo ago), licensed MIT. It adds 140 tokens to every session and 1,715 once invoked, about $0.0007 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 commands, from other repositories
e2e
使用 Playwright 对 Web UI 进行端到端测试(支持视频录制、Trace 录制、控制台/网络日志捕获).
auto-browse
Auto-browse — learn, optimize, and graduate browser operations or web data-mining workflows.
research-perplexity
Run a deep research query using Perplexity's /research mode via Playwright browser automation. This is an alternative to /export-to-council that uses Perplexity's dedicated research mode instead of multi-model council.
review
Compare a reference design against an implementation. Accepts Figma URL, image file, or browser URL as reference.
graphify
Turn your vault into a clustered knowledge graph with HTML and JSON outputs.
laravel-playwright
E2E Playwright patterns; use the laravel:e2e-playwright skill exactly as written.