Watch official sources for changes

Name the pages to watch. Get a source comparison; a failed check never reads "no change".

What an agent needs to own a weekly job ↗

Tried once · source comparison

A fresh local operator used the exported source-watch instruction with a brief selecting one official page. It reused a retained 13 October baseline, retrieved the official page on 15 October at 02:48:34 UTC and saved the comparison. The documentation connector was unavailable; official-domain web retrieval succeeded. Not scheduled. This is one local comparison, not Dot execution, unattended monitoring or a complete product audit. Source modification time, intermediate edits and web-reader cache freshness remain unknown. Full source captures remain in the private evidence vault.

Local schedules require the supported computer/app runtime. Cowork tasks that need local files or apps run locally; cloud runs need remotely accessible inputs.

Read the full instruction ↓

What you supply

A small named list of official URLs and the changes that matter
Baseline location, later check time and notification rule

What it saves

Dated baseline or comparison report
Evidence for each change
Schedule proposal or host-confirmed schedule details

Designed example · synthetic data

See the finished example

Read the complete example →Starting materials →Editable Markdown ↓

The instruction

Use the official URLs and change criteria in my brief. First save a dated baseline with each URL, access result and the relevant content or source locator. On a later authorized check, compare against that saved baseline. Do not infer a change from a failed fetch.

Save source-watch.md and the baseline/comparison evidence in the authorized location. For every source record Changed, Unchanged or Check failed. For a change include the old and new evidence, date checked, why it matters and any uncertainty about when it occurred. Retain the previous baseline and show what was compared. Stay quiet for unchanged sources unless I request routine reports; report a relevant change, failure or required decision.

This instruction alone does not create a recurring job. If I explicitly request recurring monitoring, use the host’s supported schedule controls, show the scope, cadence, time zone, storage location, notification rule and runtime requirements, and confirm the saved schedule status. If unavailable or unconfirmed, report “Not scheduled” and leave a proposal. Do not claim monitoring is active from a connected source or completed one-off run. Do not post externally or change source accounts.

Show plain text

Use the official URLs and change criteria in my brief. First save a dated baseline with each URL, access result and the relevant content or source locator. On a later authorized check, compare against that saved baseline. Do not infer a change from a failed fetch.

Save source-watch.md and the baseline/comparison evidence in the authorized location. For every source record Changed, Unchanged or Check failed. For a change include the old and new evidence, date checked, why it matters and any uncertainty about when it occurred. Retain the previous baseline and show what was compared. Stay quiet for unchanged sources unless I request routine reports; report a relevant change, failure or required decision.

This instruction alone does not create a recurring job. If I explicitly request recurring monitoring, use the host’s supported schedule controls, show the scope, cadence, time zone, storage location, notification rule and runtime requirements, and confirm the saved schedule status. If unavailable or unconfirmed, report “Not scheduled” and leave a proposal. Do not claim monitoring is active from a connected source or completed one-off run. Do not post externally or change source accounts.

Stops in this kit

A failed check (never reported as "no change")

Posting or account changes

Turning a one-off check into a schedule

Scope, cadence, storage and notification rule

Checks to run

Every source has a result, including failed checks.
Change evidence identifies both baseline and later observation.
The report distinguishes detection time from an unknown change date.
An active schedule claim includes a host-confirmed saved job and its requirements.

Open the checklist →

What has been tested

Tried once · source comparison

Checked 2026-10-13

Limits

Local schedules require the supported computer/app runtime. Cowork tasks that need local files or apps run locally; cloud runs need remotely accessible inputs.
Dot access, connected tools and schedules vary by plan, rollout and workspace.
Copying a prompt or connecting a source does not activate monitoring.
Self-hosted agents require operator-managed permissions, hosting and delivery configuration.

Optional tool setup

Using a kit

Choose a few official sources and define changes worth noticing.
Run a baseline once and retain the saved evidence.
Compare on a later run; if recurring work is wanted, separately save and confirm a schedule.

Use a kit in your tool →

ChatGPT

Paste the instruction, attach the sample files or your own, and read the stops first.
Pick Work for a multi-step task and supply the context it needs.
For reuse, follow the documented skill or plugin setup on your surface. Downloading a kit does not install it.

Check the permissions for the tools and actions you use. The kit's stops are instructions to the agent; they do not create an enforced approval setting.

ChatGPT · dated notes →

Claude

Paste the instruction, attach the sample files or your own, and read the stops first.
In Cowork, choose the authorized folder containing your inputs where that feature is available. A generic kit-to-skill import was not established by the reviewed pages.
In Claude Code, use CLAUDE.md for project guidance. AGENTS.md is read by default on supported versions (2.1.277+) only when project-path CLAUDE.md or CLAUDE.local.md does not take precedence; session and configuration exceptions apply.

Check the controls in your Claude surface. Anthropic describes approval handbacks, but the kit's stops do not enforce a product permission boundary.

Claude · dated notes →

Codex

Paste the instruction, attach the sample files or your own, and read the stops first.
For project guidance, use AGENTS.md. For a reusable skill, follow the supported setup, such as .agents/skills/<name>/SKILL.md in a repository.
Check the sandbox and approval settings before you start the job.

Codex has sandbox and approval settings. Check them before a job that touches real files.

Codex · dated notes →

Copilot

Paste the plain instruction into a Copilot chat you can access, and read the stops first.
Check which file inputs and actions your tenant's actual surface supports before supplying files.
Keep the kit as a manual prompt. The reviewed posts do not document importing it as a skill, plugin or agent.

Ask your administrator which actions, permissions and approval controls your organization has enabled. A kit's stops do not create those controls.

Copilot · dated notes →

Gemini

Paste the instruction, attach the sample files or your own, and read the stops first.
Where Skills is available, copy or import the skill using that surface's documented setup.
Gemini app and Workspace skills do not sync. Set up the skill separately on each available surface you use.

Some Gemini Tasks may require confirmation. Check current eligibility and permissions before real account work; the kit's stops are instructions, not enforced approval settings.

Gemini · dated notes →

Grok

Paste the plain instruction as a message and read the stops first.
If Bot is available to you, show it the workflow; xAI describes routine learning by following along.
Keep the instruction as plain text; we could not confirm SKILL.md support.

xAI describes approval handbacks for Grok Bot. Check what your account and connected apps allow before real actions.

Grok · dated notes →

Other agents

Paste the plain instruction into an agent that accepts text, then supply inputs through its supported route and read the stops first.
If your agent reads AGENTS.md, save the instruction there.
If the named client supports Agent Skills, use the documented skill folder location and setup for that client and version.

This varies by agent. Look up its approval settings; the kit's stops stay in the instruction wherever you paste it.

Other agents · dated notes →

Markdown · README.md

# Watch official sources for changes

Name the pages to watch. Get a source comparison; a failed check never reads "no change".

## Start

1. Edit [brief.md](source-change-watch/files/brief.md) and [input-template.md](source-change-watch/files/input-template.md).
2. Supply authorized inputs to your own file-capable agent with [instruction.txt](source-change-watch/files/instruction.txt).
3. Inspect the saved outputs using [checks.md](source-change-watch/files/checks.md) before using them. Keep original inputs.

## Inputs

- A small named list of official URLs and the changes that matter
- Baseline location, later check time and notification rule

## Outputs

- Dated baseline or comparison report
- Evidence for each change
- Schedule proposal or host-confirmed schedule details

## Steps

- Choose a few official sources and define changes worth noticing.
- Run a baseline once and retain the saved evidence.
- Compare on a later run; if recurring work is wanted, separately save and confirm a schedule.

## Instruction

Use the official URLs and change criteria in my brief. First save a dated baseline with each URL, access result and the relevant content or source locator. On a later authorized check, compare against that saved baseline. Do not infer a change from a failed fetch.

Save source-watch.md and the baseline/comparison evidence in the authorized location. For every source record Changed, Unchanged or Check failed. For a change include the old and new evidence, date checked, why it matters and any uncertainty about when it occurred. Retain the previous baseline and show what was compared. Stay quiet for unchanged sources unless I request routine reports; report a relevant change, failure or required decision.

This instruction alone does not create a recurring job. If I explicitly request recurring monitoring, use the host’s supported schedule controls, show the scope, cadence, time zone, storage location, notification rule and runtime requirements, and confirm the saved schedule status. If unavailable or unconfirmed, report “Not scheduled” and leave a proposal. Do not claim monitoring is active from a connected source or completed one-off run. Do not post externally or change source accounts.

## Decisions that need your context

- A failed check (never reported as "no change")
- Turning a one-off check into a schedule
- Scope, cadence, storage and notification rule
- Posting or account changes

## Settled by the brief

No separately annotated settled rules are supplied. Follow the instruction as written.

## Check the output

- Every source has a result, including failed checks.
- Change evidence identifies both baseline and later observation.
- The report distinguishes detection time from an unknown change date.
- An active schedule claim includes a host-confirmed saved job and its requirements.

## Limits

- Local schedules require the supported computer/app runtime. Cowork tasks that need local files or apps run locally; cloud runs need remotely accessible inputs.
- Dot access, connected tools and schedules vary by plan, rollout and workspace.
- Copying a prompt or connecting a source does not activate monitoring.
- Self-hosted agents require operator-managed permissions, hosting and delivery configuration.

## Materials

- [examples/source-change-watch/designed-example.html](source-change-watch/files/examples/source-change-watch/designed-example.html)
- [examples/source-change-watch/designed-example.md](source-change-watch/files/examples/source-change-watch/designed-example.md)
- [examples/source-change-watch/input.md](source-change-watch/files/examples/source-change-watch/input.md)
- [examples/source-change-watch/recorded-run/instruction.txt](source-change-watch/files/examples/source-change-watch/recorded-run/instruction.txt) (original recorded evidence; links retain their recorded form)
- [examples/source-change-watch/recorded-run/output/comparison.md](source-change-watch/files/examples/source-change-watch/recorded-run/output/comparison.md) (original recorded evidence; links retain their recorded form)
- [examples/source-change-watch/recorded-run/report.html](source-change-watch/files/examples/source-change-watch/recorded-run/report.html) (original recorded evidence; links retain their recorded form)
- [examples/source-change-watch/recorded-run/report.md](source-change-watch/files/examples/source-change-watch/recorded-run/report.md) (original recorded evidence; links retain their recorded form)

## Evidence

[Recorded context and file identities](source-change-watch/files/evidence.json)

Tried once · source comparison

A fresh local operator used the exported source-watch instruction with a brief selecting one official page. It reused a retained 13 October baseline, retrieved the official page on 15 October at 02:48:34 UTC and saved the comparison. The documentation connector was unavailable; official-domain web retrieval succeeded. Not scheduled. This is one local comparison, not Dot execution, unattended monitoring or a complete product audit. Source modification time, intermediate edits and web-reader cache freshness remain unknown. Full source captures remain in the private evidence vault.

Sources checked: 2026-10-13. Recorded run: 2026-10-15T02:48:34Z.

## Revision and corrections

Revision: 2.0.1. Lifecycle: active.

[Current kit](https://codexwork.com/kits/source-change-watch/) · [Changes and corrections](https://codexwork.com/changes/)

[Rights and notices](source-change-watch/files/LICENSES.md) · [Sources](source-change-watch/files/sources.md) · [Manifest](source-change-watch/files/manifest.json)

Downloading these files does not run an agent, install a skill, create a schedule, or grant file or account permissions.

Open Markdown file ↗

SKILL.md

---
name: source-change-watch
description: "Name the pages to watch. Get a source comparison; a failed check never reads \"no change\"."
---

# Watch official sources for changes

## Required inputs

- A small named list of official URLs and the changes that matter
- Baseline location, later check time and notification rule

## Expected outputs

- Dated baseline or comparison report
- Evidence for each change
- Schedule proposal or host-confirmed schedule details

## Checks

- Every source has a result, including failed checks.
- Change evidence identifies both baseline and later observation.
- The report distinguishes detection time from an unknown change date.
- An active schedule claim includes a host-confirmed saved job and its requirements.

## Limits

- Local schedules require the supported computer/app runtime. Cowork tasks that need local files or apps run locally; cloud runs need remotely accessible inputs.
- Dot access, connected tools and schedules vary by plan, rollout and workspace.
- Copying a prompt or connecting a source does not activate monitoring.
- Self-hosted agents require operator-managed permissions, hosting and delivery configuration.

Revision: 2.0.1. [Read the complete kit](files/README.md) before use. This optional format is not a claim of support by every tool.

Open SKILL.md ↗