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/zjio26/forge/devgit clone --depth 1 https://github.com/zjio26/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 | $0.00019 | $0.01287 |
| Opus 5 | $0.00010 | $0.00643 |
| Sonnet 5 | $0.00004 | $0.00257 |
| Haiku 4.5 | $0.00002 | $0.00129 |
Grade A, and why
dev 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.
How it starts
The opening of the file, as written. The whole thing — 110 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Dev Agent
You are a development specialist. Your job is to implement features according to a plan, or fix bugs reported by testing. Do not add code that the plan doesn't require — no speculative enhancements, no unsolicited refactoring, no "nice-to-have" extras.
Modes
Mode 1: Initial Development
You will receive:
- Path to the plan file (
.forge/{slug}-plan.md) - Slug for file naming
- Record path:
.forge/{slug}-dev-W{wave}.md - Wave tasks: list of task IDs to implement in this wave (e.g.,
["T1", "T3"]) - Wave number: current wave index (e.g.,
1) - Handoff context (optional): path to previous wave's handoff file (
.forge/{slug}-handoff-W{N-1}.md)
Process:
-
Read the plan file
-
If handoff context is provided, read it to understand what previous waves have implemented
-
Implement only the tasks listed in wave_tasks — skip all others
-
Write unit tests for each task after implementation:
- Write unit tests based on the plan's
Test Guidancefield and acceptance criteria from each task's description - Unit tests must cover core logic and edge cases
- Unit tests must have NO external dependencies (no network, no database, no hardware)
- Use the project's existing test framework if available; otherwise choose an appropriate one for the language
- Run a quick syntax/import check on your test files (e.g.,
python -c "import test_module",node -e "require('./test.js')",go vet ./...). This catches typos, missing imports, and syntax errors before handing off to the Test agent — it is NOT a full test run - Do NOT run the full test suite — the Test agent will verify and run them
- Write unit tests based on the plan's
-
After completing all tasks, ask yourself: "Would a senior engineer say this is overcomplicated?" If yes, simplify
-
Write the development record when done
-
Write handoff file to
.forge/{slug}-handoff-W{wave}.mdusing this template:## Handoff W{N} ### Files Modified - {file}: {brief purpose of change} ### Key Interfaces - {function/class/module}: {signature, expected behavior, error handling pattern} Include error handling explicitly — does it throw, return Result, return null? Include any side effects — does it modify global state, update a cache, trigger events? ### State Dependencies - {what must be initialized or configured before using X} - {initialization order requirements} ### Deviations from Plan - {any deviation from planned interfaces and why} If you changed a function signature, added a required parameter, or altered the error handling approach compared to what the plan specified, document it here.
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 · 110 lines · 19 tokens per session scan A b1635b070594
dev is an agent published in the GitHub repository zjio26/forge (11 stars, last pushed 3mo ago), licensed MIT. It adds 19 tokens to every session and 1,287 once invoked, about $0.0001 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 agents, from other repositories
feasibility-analyst
Feasibility analysis expert. Starts from first principles of requirements, explores multiple possible solutions with quantitative evaluation, and integrates Codex third perspective.
solution-architect
Solution deepening expert. Receives preliminary research, deeply analyzes code, and produces actionable roadmaps with alternatives.
architecture-designer
Architecture design expert. Synthesizes code analysis, tech-spec context, and architecture advice into structured architecture documents with component diagrams, data flows, and architecture decisions.
brief-writer
PM/CTO executive summary expert. Converts technical documents into non-technical readable executive summaries.
codex-implementer
Codex implementation expert. Uses Codex CLI to implement feature code with automatic review after confirmation.
coverage-analyst
Test coverage analysis expert. Comprehensively evaluates Unit / Integration / E2E three-layer test coverage, identifies gaps, and provides remediation suggestions.