audit_pre_agent_workspace.sh can exceed the 1 MiB step summary limit when toolchains are installed under RUNNER_TEMP/gh-aw/tool-cache
Maintainers usually reply within 1 day
Nobody has claimed this yet.
Assessment
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Newbie friendliness
- 70/100
Research direction
Start in actions/setup/sh/audit_pre_agent_workspace.sh, at the exclusion list for cache directories and at render_summary_tree, which writes to GITHUB_STEP_SUMMARY. The issue offers three options (exclude tool-cache, cap the summary, or summarize only counts for RUNNER_TEMP/gh-aw), so confirm the choice with a maintainer first. Done means a .NET install under RUNNER_TEMP/gh-aw/tool-cache no longer pushes the summary past 1 MiB, while /tmp/gh-aw/pre-agent-audit.txt keeps the full listing.
Written by the indexing model from the issue text.
Description
Problem
actions/setup/sh/audit_pre_agent_workspace.sh appends a rendered file tree of every scanned directory to $GITHUB_STEP_SUMMARY (render_summary_tree >> "${GITHUB_STEP_SUMMARY:-/dev/null}") with no size cap. One of the scanned directories is $RUNNER_TEMP/gh-aw. When the "Redirect tool cache and install paths for ARC/DinD" step has put the setup-dotnet and setup-node installs under $RUNNER_TEMP/gh-aw/tool-cache, that directory holds thousands of files and the rendered summary exceeds the per-step limit (1 MiB, see Adding a job summary). GitHub then fails the upload for that step and adds an error annotation.
The run itself succeeds, so this is cosmetic, but every run of such a workflow carries a red error annotation and loses that step's summary.
Evidence
One implement-issue run (gh-aw v0.89.21, Claude engine, a .NET repository with actions/setup-dotnet and actions/setup-node):
- Log:
Pre-agent audit written to /tmp/gh-aw/pre-agent-audit.txt (9992 lines), then##[error]$GITHUB_STEP_SUMMARY upload aborted, supports content up to a size of 1024k, got 1385k. - Of the 9,992 audit lines, 9,941 are under the
gh-aw temp directorysection: 5,751 undertool-cache/dotnetand 3,448 undertool-cache/node. The workspace sections add 51 lines. - A run of another workflow in a different repository without a .NET install had 741 lines in the same section and did not hit the limit.
The script already excludes common cache directories (node_modules, .cache, vendor, ...) from the listing; tool-cache is not among them.
Suggested fixes
Any one of these would avoid it:
- Exclude
tool-cache(and other installed-toolchain directories) from the$RUNNER_TEMP/gh-awscan, as with the existing cache exclusions. - Cap what is written to the step summary (for example the first N lines plus "... M more, see /tmp/gh-aw/pre-agent-audit.txt"), keeping the complete listing in the audit file.
- Write the summary tree only for the workspace and home directories, and only the counts for
$RUNNER_TEMP/gh-aw.
- Dominant language
- Go
- Stars
- 5.4k
- Forks
- 576
- Avg merge
- 7h 21m
- Merged PRs (30d)
- 816
Getting set up
Starts the project's dev container in your browser, under your own GitHub account.
- Ships a Dockerfile or Docker Compose file
- No pull request template
- Read the contributing guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from github/gh-aw
-
automation awf compiler specifications
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
Maintainers usually reply within 1 day
-
automation models
Difficulty 3/5 Half a day Newbie friendliness 22/100
Maintainers usually reply within 1 day
-
agentic-workflows
Difficulty 4/5 3-5 days Newbie friendliness 12/100
Maintainers usually reply within 1 day
-
agentic-workflows
Difficulty 4/5 1-2 days Newbie friendliness 12/100
Maintainers usually reply within 1 day
-
agentic-workflows
Difficulty 3/5 Half a day Newbie friendliness 25/100
Maintainers usually reply within 1 day
Similar issues
-
Action Required: Fix Renovate ConfigurationPossibly taken @thompson-tomo claimed this today. Open
Difficulty 1/5 Under an hour Newbie friendliness 78/100
open-telemetry/opentelemetry-collector#16170 ·
Maintainers usually reply within 1 day
-
state:triage-needed
Difficulty 2/5 1-3 hours Newbie friendliness 65/100
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 80/100
puppe1990/amarra-cais#342 ·
Maintainers usually reply within 1 day
-
replay: a run cancelled during environment setup is logged as an ERROR (Instrument drops the cancellation)Possibly taken A pull request linked to this issue is open or already merged. Open
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 70/100
Maintainers usually reply within 1 day