source-command-latch-pm

A guided demonstration for recording one decision that the project has ruled out. It then shows how latch can stop a later agent from proposing that decision again.

In plain words
What is it for?
Use it to capture a rejected library, framework, service, architecture, or workflow in latch’s knowledge base, after reviewing and confirming the entry.
Why use it?
It preserves an important judgment so future coding agents do not quietly revisit the same rejected approach.

Skill for Claude CodeCodex

Install

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.

agentmods
npx agentmods add skills/open-latch/latch/source-command-latch-pm
Any agent
npx skills add open-latch/latch --skill source-command-latch-pm
Clone the repo
git clone --depth 1 https://github.com/open-latch/latch

Made for: Claude Code, Codex.

Per session 65 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 429 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00065 $0.00429
Opus 5 $0.00032 $0.00215
Sonnet 5 $0.00013 $0.00086
Haiku 4.5 $0.00006 $0.00043

Measured 2d ago against content hash 9d4c4b991b14, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

source-command-latch-pm 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.

.agents/skills/source-command-latch-pm/SKILL.md · 39 lines

What it actually says

source-command-latch-pm

Use this skill when the user wants the latch PM demo/workflow in Codex: capture one concrete ruled-out decision and show latch's decision-continuity guardrail.

Workflow

  1. Read what latch already knows with the available latch KB read tools (latch_recent/kb_recent, priority listing, and search as needed). Do not re-ask what is already captured.

  2. Tell the user in one line that you will help seed one decision latch can use to stop a future agent from quietly undoing their judgment.

  3. Ask for one concrete approach they already ruled out and why. Good fuel is a specific library, framework, service, architecture, or workflow that another agent might plausibly propose again.

  4. If the answer is vague, ask at most one sharpening question. If it is still too vague, fall back to a lightweight project snapshot rather than turning this into an interrogation.

  5. Confirm before writing. Render the proposed node with kind, title, body, and any links. On accept/edit, write it with latch_insert/kb_insert as a decision in staging status and link it to the current workstream when one is known. Skip means no write.

  6. After the decision is saved, invite the user to trigger the trip-wire by asking for the rejected path. For the strongest proof, suggest trying it in a fresh session too. When they do, run latch search/gate, cite the stored decision, and redirect with the saved reason.

Keep the payoff focused on decision continuity, not generic project summaries.

Files

What ships with it

1 file 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.

Changes

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.

  1. 2d ago First seen · 39 lines · 65 tokens per session scan A 9d4c4b991b14

Subscribe to this mod's changes

source-command-latch-pm is a skill published in the GitHub repository open-latch/latch (9 stars, last pushed 2d ago), licensed Apache-2.0. It adds 65 tokens to every session and 429 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.