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 skills add ntorga/agent-starter-kit --skill architect-impl-groundinggit clone --depth 1 https://github.com/ntorga/agent-starter-kitWrote 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/skills/ntorga/agent-starter-kit/architect-impl-grounding)<a href="https://agentmods.dev/skills/ntorga/agent-starter-kit/architect-impl-grounding"><img src="https://agentmods.dev/badge/skills/ntorga/agent-starter-kit/architect-impl-grounding/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/ntorga/agent-starter-kit/architect-impl-grounding"><img src="https://agentmods.dev/badge/skills/ntorga/agent-starter-kit/architect-impl-grounding.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00047 | $0.00878 |
| Opus 5 | $0.00023 | $0.00439 |
| Sonnet 5 | $0.00009 | $0.00176 |
| Haiku 4.5 | $0.00005 | $0.00088 |
Grade A, and why
architect-impl-grounding 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 — 75 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Purpose
The grill settles the decisions. This skill makes them concrete. It grounds the plan artifacts in the codebase and annotates impl.md with the details the coder implements from. The refresh procedure handles what lands: it verifies the shipped epic's discoveries, classifies them, and re-grounds the next epic against the codebase as it now stands.
Procedure
- Ground — the grill is done. Annotate
impl.md. - Refresh — an epic has landed. Classify its discoveries, then re-ground the next epic.
- Ground mode. Annotate
impl.md: a. Read the plan artifacts:plan.md— acceptance criteria.arch.md— directory structure, layer separation, reference projects.impl.md— epics grouped at one day of work. b. Ground the artifacts in the codebase:- Read
.context.mdfiles in affected directories. - Read
docs/FEATURE-MAP.mdif it exists. - Read the project's architecture skill, if one exists.
- Run
lson affected directories. - Verify assumptions — check library docs for capability claims, inspect handler code for API contracts, validate environment references against config files.
c. Annotate
impl.mdwith the annotation set for each epic. d. Respect current layer boundaries and dependency direction (follows:rules/code/general.md). For each file in the epic, verify its imports do not violate dependency direction. e. Write the annotatedimpl.mdback to disk. The coder works from this file.
- Refresh mode. Handle the landed epic's discoveries, then re-ground:
a. Read
impl.mdand the coder's handoff for the epic that landed. b. Verify each discovery against the changed files. Usegit diff --statto see what shipped, then read the files the discovery claims to affect. Discard a discovery that does not match the implementation — note the mismatch. c. Classify each verified discovery:- adjust — changes how the remaining work is built, not what the user wants. Fold it into the remaining epics inline. No user gate.
- escalate — contradicts a settled decision, or reveals a decision the user should make. Flag it for the Maestro. Do not guess.
d. Re-ground the next pending epic with step 1 against the current codebase. Verify its file paths, method signatures, and reference files. Re-annotate what drifted.
e. Write the updated
impl.mdback to disk.
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 · 75 lines · 47 tokens per session scan A 43b4cc74fe9e
architect-impl-grounding is a skill published in the GitHub repository ntorga/agent-starter-kit (142 stars, last pushed 2d ago), licensed MIT. It adds 47 tokens to every session and 878 once invoked, about $0.0002 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 2026-09-13.
Other skills, from other repositories
board
Read and write the workspace BOARD.md (shared task board for agents + Val). Wraps bin/board, a small Python CLI that preserves the markdown schema and appends to the activity log automatically. Use this whenever an agent proposes, claims, or finishes work that should be visible to the rest of the roster.
squads-cli
Squads CLI — operate your AI workforce through the loop; intent in, reviewed work out. TRIGGER when running squads, dispatching agents, triaging the inbox, reading/writing memory, recording feedback, or checking status/usage.
gh
You have access to gh (GitHub CLI) for repository and project management.
pm
Project Manager skill. One orchestrator session drives eight subagent roles -- planner, plan-reviewer, doer, reviewer, deployer, integ-test-runner, ci-watcher, harvester -- through cycles of Plan, Develop, Test, and Harvest across one or more parallel tracks in isolated git worktrees, running each task on a…
auto-sprint-args
How to construct and pass correct arguments to the auto-sprint workflow -- the required issues+branch fields, optional goal/maxcycles/basebranch/requirementsFile, JSON-object (not string) shape, and common launch mistakes. Trigger when preparing to launch or invoke auto-sprint.
fleet-sprint-cli
How to launch and supervise apra-fleet's autonomous sprint workflow from the CLI (apra-fleet workflow fleet-sprint) -- the --issue/--members/--branch/--base flag contract, the preconditions to check first, how to launch it as a detached process, and how to watch it via its dashboard. Trigger before running apra-fleet…