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 skills/sandsower/beislid/implementnpx skills add sandsower/beislid --skill implementgit clone --depth 1 https://github.com/sandsower/beislidWhat 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 | $0.00045 | $0.02877 |
| Opus 5 | $0.00023 | $0.01438 |
| Sonnet 5 | $0.00009 | $0.00575 |
| Haiku 4.5 | $0.00005 | $0.00288 |
Grade A, and why
implement 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 — 134 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Implementation Plan
Creates a structured plan, tracks it with the host agent's todo/task mechanism, and executes with TDD as the default rhythm.
When beislid:agent_isolation is configured, load workspace placement plus the current host adapter.
Only with that block present, run ensure_orchestrator_workspace before the first repo write and place_mutating_delegate before each mutating dispatch.
When the current host is Codex and a subagent dispatch is planned, load Codex delegate context just before dispatch even when agent isolation is not configured.
Also load it when a smoke fixture explicitly requests protocol proof.
If the handoff includes an explicit design artifact path from blueprint, read it as your primary input; design artifacts are checkpoint-compatible state seeds for implementation planning. Otherwise, if the matching blueprint_approved latest pointer entry for the current ticket/branch resolves to a readable artifact, read that artifact as your primary input; if not, use the workflow-configured artifact path template when present, then look for a matching design artifact in plans/ using the ticket/feature slug when known (for example, plans/<feature>-design.md from blueprint). If exactly one match exists, read it as your primary input. If multiple candidates remain, ask the user to choose the artifact path. Only fall back to conversation context when no design artifact is available.
If the handoff includes an approved execution-envelope-v0, treat its autonomy boundary as execution context for this agent or an external runner: allow lists pre-approved work, ask creates human approval stops before continuing, and deny is out of scope. The envelope may carry Beislið ProcessProvider semantics such as selected contracts, slices, proofs, gates, guides, model hints, action-policy boundaries, and capability/probe status. Do not reinterpret an envelope as a runtime service, parser, durable run store, Rondo adapter contract, or replacement for Work Contract requirements.
If the user is resuming with phrases such as continue this ticket, continue implementation, or continue from checkpoint, read .beislid/checkpoints/latest.json when present. Prefer an implementation_plan_created checkpoint matching the current branch and ticket ID when known; otherwise fall back to a matching kickoff_context_ready checkpoint or ask the user to choose among matching latest entries. Read the referenced checkpoint artifact as primary context before falling back to conversation context. Missing, unreadable, or malformed latest pointers are non-blocking; warn when malformed, then ignore the pointer and fall back to a matching checkpoint artifact or conversation context. If durable run-ledger state is available, beislid run-ledger resume --flow implement --ticket-id <id> --branch <branch> may identify the latest running/interrupted/failed external run, but checkpoint artifacts/design artifacts remain the primary content seed for implementation planning.
Action-risk decisions follow action-policy-protocol.md; read it before checkpoint writes, workspace edits, dependency installs, local git operations, or configured side-effect hooks. For durable run evidence when ticket/branch context is known, follow the nopal-seam-protocol.md fallback ladder. With mode: off, use beislid run-ledger init --skill implement --flow implement. Otherwise probe(nopal_seam) and, only when the probe is ok and capabilities[] contains ledger, call nopal ledger init --skill implement --flow implement --json. Under mode: prefer, an unavailable probe, missing ledger capability, or failed Nopal call warns and falls back to beislid run-ledger init; under mode: require, the same condition blocks instead of falling back. Record transcript-safe events for plan creation, task-batch starts/completions, verification results, and interruptions. When a workflow checkpoint artifact is written, add a checkpoint through the same seam (nopal ledger checkpoint --json or beislid run-ledger checkpoint) with --run-id <run_id> --flow implement --name implementation_plan_created --resume-hint <resume_hint> so the resume hint stays attached to the run. Ledger failures warn only on a fallback-eligible Beislid path and never replace task tracking, verification, or checkpoint artifact behavior.
What ships with it
9 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.
- action-policy-protocol.md 40 B
- artifact-templates.md 36 B
- codex-delegate-context.md 2.8 KB
- nopal-seam-protocol.md 37 B
- workspace-placement-claude.md 1.9 KB
- workspace-placement-codex.md 2.6 KB
- workspace-placement-generic.md 1.2 KB
- workspace-placement-pi.md 1.8 KB
- workspace-placement-protocol.md 3.7 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.
- 2d ago First seen · 134 lines · 45 tokens per session scan A 617e53c9ff8c
implement is a skill published in the GitHub repository sandsower/beislid (10 stars, last pushed 2d ago), licensed MIT. It adds 45 tokens to every session and 2,877 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 2026-08-31.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
babysit-pr
Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…
imagegen
Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…