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/dormstern/forge/buildgit clone --depth 1 https://github.com/dormstern/forgeWhat 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 | $0.00058 | $0.00883 |
| Opus 5 | $0.00029 | $0.00441 |
| Sonnet 5 | $0.00012 | $0.00177 |
| Haiku 4.5 | $0.00006 | $0.00088 |
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 — 61 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/forge:build — Ralph Loop (PM mode)
Execution mode: direct. This command has explicit steps below — execute them in order. Do not enter plan mode, do not write a plan file, do not call ExitPlanMode. The framework already did the planning; you just run the steps.
You execute the SHAPE-approved plan, feature by feature. Fresh context per feature — never carry context across feature boundaries.
Pre-conditions
releases.json.current_cycle.approved_planexists. If not, instruct the user to run/forge:shapefirst.- Optional:
/forge:build <feature-id>to build a specific feature. Default is wave-aware selection.
Behavior
Startup Protocol
- Run the project's full test suite. Must be green. If not, fix the green-on-master invariant before building anything new.
- Verify
architecture.mdis current (post-last-build). - Confirm clean working tree.
- Read
progress.mdtop sections (Cross-Cutting Patterns, Active Gotchas) for ~1k tokens of context.
Ralph Loop (per feature)
- SELECT the next feature: current wave first, all hard
depends_onmet, debt before dependent new features, removal features last in wave. - Mark
status="building"infeatures.json. Print:--- CONTEXT BOUNDARY -- starting feature: {id} --- - Dispatch test-writer to produce 3–7 acceptance tests. For API features (if
contract_mode = "full"), contract-compliance tests come first. - Dispatch builder with fresh context: only the test file, the relevant
architecture.mdsection (~500 tokens), andprogress.mdtop sections. Max 3 resets. - EVALUATE:
- If tests pass: commit per autonomy tier (T1 auto, T2 checkpoint before commit, T3 checkpoint before tests AND before commit). Set
status="done". Calculatequality_score(rubric below). Append learnings toprogress.md. - If failed after 3 resets: set
status="blocked", log the blocker, append learnings, continue with the next feature.
- If tests pass: commit per autonomy tier (T1 auto, T2 checkpoint before commit, T3 checkpoint before tests AND before commit). Set
- Drift check every 5 features. Dispatch drift-detector. If recommendation is DISCUSS / PIVOT / CONTRACT_HALT, PAUSE and notify the user. Do not advance.
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 · 61 lines · 58 tokens per session scan A 50ae82a1547d
build is a command published in the GitHub repository dormstern/forge (6 stars, last pushed 3mo ago), licensed MIT. It adds 58 tokens to every session and 883 once invoked, about $0.0003 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 2026-08-31.
Other commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.