Propose a patch
A fix the maintainer can act on, with the evidence attached. It opens under your GitHub account in your browser; KillerSkills holds no token and never modifies the package. The fix is verified the only way we verify anything: the next crawl re-runs the new version in the honeypot. 68 · Caution
Evidence
The sandbox observed nothing beyond the static findings; the proposal cites those.
Diff · removes 9 caught lines
Download .patchdiff --git a/src/terse/cli.py b/src/terse/cli.py --- a/src/terse/cli.py +++ b/src/terse/cli.py @@ -1794,1 +1794,0 @@ -# A URL-shaped payload (`-H https://u:pw@h`, `-H tcp://...`) is a URL, not a header. @@ -1842,1 +1841,0 @@ -"""`url` with any userinfo and query string masked: `https://u:pw@h/mcp?token=x` is a @@ -1844,1 +1842,0 @@ -(`https://u:pw@[::1/x`, `https://u:pw@/x`), goes through `_redact_url_text` instead @@ -1958,1 +1955,0 @@ -# `--url=https://u:pw@h`, `DATABASE_URL=postgres://u:pw@db`, `--env=K=v`. @@ -2657,1 +2653,0 @@ -help="behavioral eval (#295): does a real tool-calling model's downstream " diff --git a/src/terse/stats.py b/src/terse/stats.py --- a/src/terse/stats.py +++ b/src/terse/stats.py @@ -74,1 +74,0 @@ -# `https://u:pw@[::1/x`: urlparse raises, and this runs in mcp-status, stats @@ -79,1 +78,0 @@ -# `https://u:pw@/x` has userinfo and no host: nothing names it, and `/x` would. diff --git a/src/terse/transforms.py b/src/terse/transforms.py --- a/src/terse/transforms.py +++ b/src/terse/transforms.py @@ -1095,1 +1095,0 @@ -value). Shared by the proxy (what ships) and the fluency-for-diff eval (what's diff --git a/src/terse/fluency/harnesses.py b/src/terse/fluency/harnesses.py --- a/src/terse/fluency/harnesses.py +++ b/src/terse/fluency/harnesses.py @@ -583,1 +583,0 @@ -build_terminal_diff_report be reused UNCHANGED for text-diff-eval (a deliberate
Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 0.51.0. If the behavior is intended, declare it (host under permissions.network, a line in the README) instead of removing it; the next crawl scores declarations.
Issue / PR body
Title: terse-mcp: URL with embedded credentials (KillerSkills verdict caution 68)
## What the honeypot observed KillerSkills ran **[email protected]** in an isolated sandbox (decoy credentials, recording proxy, syscall trace; engine 1.9.2) and scored it **CAUTION 68/100**. Summary: no network, touches nothing outside the project, writes no memory. - See the findings below. Full trace and findings: https://killerskills.ai/skills/pypi/terse-mcp/0.51.0 ## Findings with a location - `src/terse/cli.py:1844` · URL with embedded credentials — src/terse/cli.py:1844 · URL with embedded credentials — may be phishing or credential leaking - `src/terse/stats.py:79` · URL with embedded credentials — src/terse/stats.py:79 · URL with embedded credentials — may be phishing or credential leaking - `src/terse/stats.py:74` · URL with embedded credentials — src/terse/stats.py:74 · URL with embedded credentials — may be phishing or credential leaking - `scripts/bench/corpus/gh_commits.json:1` · URL with embedded credentials — scripts/bench/corpus/gh_commits.json:1 · URL with embedded credentials — may be phishing or credential leaking - `src/terse/cli.py:1842` · URL with embedded credentials — src/terse/cli.py:1842 · URL with embedded credentials — may be phishing or credential leaking - `scripts/bench/corpus/gh_repo_single.json:1` · URL with embedded credentials — scripts/bench/corpus/gh_repo_single.json:1 · URL with embedded credentials — may be phishing or credential leaking - `src/terse/cli.py:1794` · URL with embedded credentials — src/terse/cli.py:1794 · URL with embedded credentials — may be phishing or credential leaking - `src/terse/cli.py:1958` · URL with embedded credentials — src/terse/cli.py:1958 · URL with embedded credentials — may be phishing or credential leaking - `src/terse/cli.py:2657` · Python exec or eval of computed code — src/terse/cli.py:2657 · Python exec or eval of computed code - `src/terse/transforms.py:1095` · Python exec or eval of computed code — src/terse/transforms.py:1095 · Python exec or eval of computed code - `src/terse/fluency/harnesses.py:583` · Python exec or eval of computed code — src/terse/fluency/harnesses.py:583 · Python exec or eval of computed code ## Proposed change The attached patch removes the 9 lines above. If the behavior is intended, the fix is to declare it instead: list the host under `permissions.network` (or the MCP manifest) and say in the README why the package needs it; the next crawl will re-score it with the declaration taken into account. ## How this will be verified Nothing is "verified safe" by assertion. When a new version is published, KillerSkills re-runs it in the honeypot and the verdict page updates; you can trigger it early from the verdict page ("Re-scan after I publish a fix"). Reply here or at [email protected] if you believe the verdict is wrong; disputes are reviewed by a person and the outcome is published on the verdict page. _Proposed through killerskills.ai/threats by a user of the package, under their own GitHub account. The evidence is public; the package was never modified by KillerSkills._
# or from a terminal, under your own account: gh repo fork inth3shadows/terse --clone && cd terse curl -s https://killerskills.ai/skills/pypi/terse-mcp/0.51.0/patch.diff | git apply --unidiff-zero git checkout -b killerskills/0.51.0 && git commit -am "terse-mcp: URL with embedded credentials" gh pr create --fill --body-file <(curl -s https://killerskills.ai/skills/pypi/terse-mcp/0.51.0/patch.md)
We count proposals opened, not merged: a merge shows up as a new version that the crawl re-runs. Disputes: the maintainer can contest the verdict from its page; a person reviews it and the outcome is published.