CategoryWeightCategory scoreContribution

Securityprompt, exec, net, exfil, eval

35%

30.8 pts

Supply chainhash, typosquat, maintainer, lockfile

20%

100

20.0 pts

Maintenancestaleness, pinning, CI

15%

100

15.0 pts

TransparencySKILL.md, perms, README

15%

100

15.0 pts

Communityinstalls, verify, response

15%

100

15.0 pts

Findings & checks · 1 flagged

Securityscore 88 · 1 finding

MEDIUMInstruction telling the agent not to ask for approvalSS-SKILL-INJECT-DONT-ASK-01 · Prompt injection · skills/deploy-log-doctor/SKILL.md

MEDIUM — it fires on intent; the real damage depends on the host agent's own approval-gating.

Why it matters

The text {match} tells the agent to skip the normal "ask the user first" gate. Used adversarially it removes the human-in-the-loop check before destructive or sensitive actions, turning a normally-gated agent into a fire-and-forget executor.

The exact value spotted

excerptskills/deploy-log-doctor/SKILL.md· markdown

1---

2name: deploy-log-doctor

3description: Diagnose failed build and deploy logs from Vercel, Netlify, Cloudflare Pages, o

… (108 chars elided on L3)

4---

Occurrences

1 occurrence · at L3

How to fix

Remove the approval-skipping instruction, or scope it narrowly to a specific safe, reversible action.

Delete blanket "don't ask / no need to confirm" directives from the skill.
If the skill is a genuine autonomous job, restrict the opt-out to a named non-destructive action rather than all actions.

Framework references

OWASPLLM01 ATLASAML.T0051

Trace & refs

ruleSS-SKILL-INJECT-DONT-ASK-01sha256ce5204b52388d24f…rubric 365aacaView on GitHub

Supply chainscore 100 · 0 findings

✓—All supply chain checks passedNo findings in this category for the latest scan.pass

Maintenancescore 100 · 0 findings

✓—All maintenance checks passedNo findings in this category for the latest scan.pass

Transparencyscore 100 · 0 findings

✓—All transparency checks passedNo findings in this category for the latest scan.pass

Communityscore 100 · 0 findings

✓—All community checks passedNo findings in this category for the latest scan.pass

The primary manifest — the file an agent reads to learn what this artifact does.

skills/deploy-log-doctor/SKILL.md · 1 file

skills/deploy-log-doctor/SKILL.md1.6 KB · Markdown

name: deploy-log-doctor

description: Diagnose failed build and deploy logs from Vercel, Netlify, Cloudflare Pages, or GitHub Actions. Use this whenever the user pastes a build log, says a deploy failed, mentions 'build error', 'deployment failed', 'works locally but not in production', or shows any CI output containing an error — even if they don't ask for a diagnosis explicitly. Runs fully offline — no API keys, no network, no credentials.

deploy-log-doctor

Part of shipsafe — offline deploy-safety skills. Every script is stdlib-only Python 3.8+; nothing leaves the machine.

Run the classifier on the log first, then interpret:

python3 scripts/diagnose_log.py <logfile>        # or pipe: cat log | python3 scripts/diagnose_log.py -
python3 scripts/diagnose_log.py <logfile> --json # structured output

The script matches 11 curated failure signatures (missing env vars, case-sensitive imports, Node version mismatches, OOM, lockfile drift, peer-dep conflicts, Prisma generate, and more) and emits the root cause plus numbered fix steps.

Workflow:

Save the user's log to a file (or pipe it) and run the script.
If it returns a diagnosis, walk the user through the fix steps, adapting them to the user's actual framework and platform.
If it returns nothing (exit 2), the log is probably truncated above the real error — ask for the section around the first ERROR line and rerun.
After fixing, suggest the rollback-readiness skill to catch the next failure class before it happens.

Exit codes: 0 diagnosed, 2 no signature matched.

All paths below are relative to this skill's directory (deploy-log-doctor/).

deploy-log-doctor— agent skill

Is deploy-log-doctor safe to install?

deploy-log-doctor

Audit the pieces. Scan the whole. Decide.