marquetools/Strip ANSI & Unicode Threats

Protects your repo from known ANSI threats and unicode homograph threats. Removes ANSI pollution from Issue and PR comments.

View on GitHub

Trust Signals

Scorecard Score: not yet scored
Maintenance Recency: Activelast commit Jun 15, 2026
License: Apache 2.0

Pinned Snippet

workflow.ymlSHA-pinned

uses: marquetools/strip-ansi-action@7699bd9f8264d6c5a8d9cae15750e5e673fd2dd6 # v1.0.6

tags can be moved; commit SHAs can't. why a SHA?

Inputs

name	description	required	default
files	Newline- or space-separated list of file paths to scan. Typically populated from the output of a changed-files step or a similar action. Leave empty when only scanning PR or Issue comments.	no	""
on-threat	Action to take when an echoback attack vector is detected. Accepted values: fail (default), strip, warn. fail — mark files, emit ::error annotations, and exit non-zero after all files are processed. strip — remove threats from output and continue; sets the threat-detected output to true. warn — emit ::warning annotations and continue; sets the threat-detected output to true.	no	fail
unicode-map	Reserved for a future release of distill-strip-ansi that will support Unicode normalization map selection. Has no effect with the current version.	no	""
no-unicode-map	Reserved for a future release of distill-strip-ansi that will support Unicode normalization map selection. Has no effect with the current version.	no	""
preset	ANSI filter preset controlling which escape sequences survive. Accepted values: dumb, color, sanitize (default), tmux, xterm, full. dumb — strip everything. color — keep SGR colors/styles only. sanitize — safe default; keeps safe OSC (titles, links), strips echoback vectors. tmux — all CSI and Fe sequences. xterm — all OSC sequences. full — pass everything through.	no	sanitize
version	Version of distill-strip-ansi to install (must match a published crates.io release).	no	0.6.1
clean-pr-comments	When true, fetches all comments on the pull request that triggered this workflow (both discussion comments and review comments) and scans them for ANSI/Unicode threats. Requires a token with pull-requests:read (and pull-requests:write when on-threat=strip). Only meaningful when the workflow is triggered by a pull_request or pull_request_review_comment event.	no	false
clean-issue-comments	When true, fetches all comments on the issue that triggered this workflow and scans them for ANSI/Unicode threats. Requires a token with issues:read (and issues:write when on-threat=strip). Only meaningful when the workflow is triggered by an issues or issue_comment event.	no	false
github-token	GitHub token used to read and (when on-threat=strip) update PR/Issue comments. Defaults to the built-in github.token. Only used when clean-pr-comments or clean-issue-comments is true.	no	""

Outputs

name	description
results	JSON array of per-file scan results. Schema: [{"file":"<path>","status":"clean\|stripped\|threat","output":"<stripped content>"}] The output field is empty for files with detected threats unless on-threat=strip. Note: large files may be truncated in this field; check runner logs for full output.
threat-detected	'true' if any scanned file contained an echoback attack vector, 'false' otherwise.
files-with-threats	Newline-separated list of file paths that contained detected threats.
comment-results	JSON array of per-comment scan results (populated when clean-pr-comments or clean-issue-comments is true). Schema: [{"id":"<id>","type":"pr_comment\|review_comment\|issue_comment","url":"<html_url>","status":"clean\|stripped\|threat","output":"<stripped content>"}]
comment-threat-detected	'true' if any scanned comment contained an echoback attack vector, 'false' otherwise. Only set when clean-pr-comments or clean-issue-comments is true.
comments-with-threats	Newline-separated list of comment HTML URLs that contained detected threats. Only set when clean-pr-comments or clean-issue-comments is true.