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 ndisisnd/msg --skill enggit clone --depth 1 https://github.com/ndisisnd/msgWrote 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/ndisisnd/msg/eng)<a href="https://agentmods.dev/skills/ndisisnd/msg/eng"><img src="https://agentmods.dev/badge/skills/ndisisnd/msg/eng/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/ndisisnd/msg/eng"><img src="https://agentmods.dev/badge/skills/ndisisnd/msg/eng.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 | $0.00130 | $0.01656 |
| Opus 5.5 | $0.00052 | $0.00662 |
| Sonnet 5.5 | $0.00026 | $0.00331 |
| Haiku 4.5 | $0.00013 | $0.00166 |
Grade A, and why
eng 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 1mo 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 — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.
eng
Platform-agnostic engineering agent with three modes — --plan, --build, --review —
each a distinct protocol in its own ref file, selected by the invocation flag. This file
routes; the shared spine that runs for every mode is refs/protocol-core.md (Steps 0–6),
and it never runs a mode's work itself.
Routing (Step 0 — full rules in refs/protocol-core.md)
Read the invocation flag and load exactly one mode protocol:
| Flag | Read |
|---|---|
--plan |
refs/plan/protocol.md |
--plan report=<path> |
refs/plan/fix-plan.md (instead of protocol.md) |
--build |
refs/build/protocol.md — unless an orchestrator spawned you as a build packet, in which case read refs/build/protocol-packet.md instead and stop; it is the whole leaf contract |
--build report=<path> |
refs/build/fix-build.md (instead of protocol.md) |
--review |
refs/review/protocol.md |
Natural-language triggers for --review: "review this change", "review my diff",
"code review this branch", "adversarial review", "look for bugs in what I just wrote".
Anything asking for style, naming or standards is /cook, not this mode.
--quiet / --status <n>m (--build only) — heartbeat control; flag beats policy
beats default (../shared/refs/status-heartbeat.md).
Exactly one mode flag must be present — zero, several, or the retired --todo hard-fail
with the exact messages in refs/protocol-core.md § Step 0. Then read the active mode
file fully and run the spine end-to-end:
The shared spine (refs/protocol-core.md)
- Step 1 — Input validation. PRD/exec-table source (mode flag +
prd-path+rows+agent, hard-refuse on any missing) or thereport=<path>alternate; more than one source is a hard failure.--reviewis exempt from Steps 1–4 — its input is the working diff; go straight to Step 5. - Step 2 — Pre-flight, one pass. Orchestrated fast path works from injected scoped excerpts (full PRD only as escape hatch); standalone path reads the full PRD, verifies row ownership mechanically via
script-prd-digest.py --verify-rows, and reads devkit + the row-relevant codebase in parallel. - Step 3 — Summary + approval gate. 3–4 line summary, then
AskUserQuestion(Yes / needs correction / follow-up → the Step 3a interview). Never proceed without an explicit "Yes, proceed" (pre-approved under an autonomy contract). - Step 4 — Coding standards (
--buildonly). Injected standards payload wins; else derive the/cookflag set and consult the persistent standards cache (script-standards-cache.py) before calling/cook(v5.6.5;../shared/refs/session-cache.md§ Consumers). - Step 5 — Run the active mode's work per its ref, then the closing message (
../shared/refs/closing-message.md) as the last chat output, doctor-logging harness incidents (../shared/refs/doctor-logging.md). - Step 6 — Scope enforcement, continuous. Only the assigned rows; unresolvable ambiguity is surfaced as a gap (plan/review) or blocks and asks (build) — never resolved by assumption.
What ships with it
11 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- refs/build/fix-build-orchestrated.md 11 KB
- refs/build/fix-build.md 9.6 KB
- refs/build/protocol-build-debug.md 1.7 KB
- refs/build/protocol-packet.md 13 KB
- refs/build/protocol.md 32 KB
- refs/plan/fix-plan.md 11 KB
- refs/plan/protocol.md 9.8 KB
- refs/plan/template-eng-plan.md 15 KB
- refs/plan/template-todo.md 7.2 KB
- refs/protocol-core.md 12 KB
- refs/review/protocol.md 8.5 KB
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.
- 1mo ago First seen · 63 lines · 130 tokens per session scan A adb666015f72
eng is a skill published in the GitHub repository ndisisnd/msg (6 stars, last pushed 2d ago), licensed MIT. It adds 130 tokens to every session and 1,656 once invoked, about $0.0005 per session on Opus 5.5. 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 skills, from other repositories
review-comments
Execute a feature-level review-comments cycle — aggregate unaddressed PR comments across every Feature.Repos PR, address or dismiss each one, and emit the standard handoff.
review-implementation-functionality-evidence
Implementation review Functionality/Evidence axis - audits plan satisfaction and verification evidence.
review-implementation-craft
Implementation review Craft axis - audits intrinsic code quality only.
review-implementation-cleanliness
Implementation review Cleanliness axis - audits change-set hygiene and pushability.
final-fix
Final-review fix pass that addresses requested changes and refreshes invalidated evidence.
code-reviewer
Automatic code quality and best practices analysis. Use proactively when files are modified, saved, or committed. Analyzes code style, patterns, potential bugs, and security basics. Triggers on file changes, git diff, code edits, quality mentions.