docs(agents): clarify sandbox escalation for progress

[kdy1]

Summary\n- add a general rule in AGENTS.md to avoid stalling work due to sandbox constraints\n- instruct agents to request escalation promptly when sandbox restrictions block required work\n\n## Testing\n- not run (documentation-only change)

[changeset-bot]

⚠️ No Changeset found

Latest commit: 82ba7e4

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

[claude]

PR Review: docs(agents): clarify sandbox escalation for progress

Summary: This is a 1-line documentation-only change to AGENTS.md, adding a rule instructing AI agents not to stall on sandbox constraints and to request escalation instead.

Feedback

The change looks fine overall. A few minor observations:

1. Clarity: The phrase "request escalation promptly and continue" is slightly ambiguous — it could be read as "request escalation and then continue working without waiting for a response," which might lead an agent to proceed with incomplete or incorrect assumptions. Consider rephrasing to clarify whether the agent should wait for the escalation to be resolved or continue with a workaround. For example:

   • "...request escalation promptly and continue with any unblocked tasks in the meantime."
   • "...request escalation promptly rather than silently stalling."

2. Placement: The rule fits reasonably in the "General rule" section, though it's more of a workflow/process instruction than a coding guideline. This is fine for now given the section already mixes coding and workflow rules.

3. No code changes / no test impact: Documentation-only, so no concerns about bugs, performance, or test coverage.

Verdict: LGTM with optional wording refinement. The intent is clear and useful — preventing agents from silently getting stuck is a good guardrail.

[Copilot]

Pull request overview

This PR adds a guideline to the AGENTS.md file instructing AI agents to proactively request escalation when sandbox constraints block required work, rather than allowing progress to stall.

Changes:

• Added a new general rule about handling sandbox constraints and requesting escalation

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The CLAUDE.md file appears to be an auto-generated or aggregated compilation of AGENTS.md files from throughout the codebase (as indicated by the "Source: AGENTS.md" markers at line 7 and similar lines throughout). However, CLAUDE.md is not being updated in this PR to reflect the new sandbox escalation guideline.

If CLAUDE.md is meant to be auto-generated, please verify that the generation process will run before this PR is merged. If it's manually maintained, it should be updated to include the new guideline in the appropriate section.

[github-actions]

Binary Sizes

File	Size
swc.linux-x64-gnu.node	28M (28572744 bytes)

Commit: 31b03ed

[codspeed-hq]

Merging this PR will not alter performance

✅ 184 untouched benchmarks

---

_{Comparing kdy1/agents-sandbox-progress-rule (82ba7e4) with main (4f35904)¹}

Footnotes

1. No successful run was found on main (85e88af) during the generation of this report, so 4f35904 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩