Tayhlow8/Caixa-D-gua-website-2026 / .agents/skills/gstack
gstack
FlaggedFast headless browser for QA testing and site dogfooding. Navigate pages, interact with elements, verify state, diff before/after, take annotated screenshots, test responsive layouts, forms, uploads, dialogs, and capture bug evidence. Use when asked to open or test a site, verify a deployment, dogfood a user flow, or file a bug with screenshots. (gstack)
Security report
FlaggedStatic analysis of one exact version. It reduces risk and shows evidence; it does not certify that a skill is safe.
- Commit
42341823ac021e7231a46e68bc0d329c1cf16577- Content hash
378a90b5f0a75cd66a7a228920eaaa7604c6412bf7b449fe7c2400f16b6b79d2- Scanned
- 2026-11-17
- Scanner version
- skillxone-scan-1.0.0
- Policy version
- 0.1-draft
- Decision
- 2 open critical or high findings.
- Verification expires
- Not applicable
Findings (2 high, 8 medium, 12 low, 7 info)
The skill refers to a token store or the system keychain, a location that holds private data or credentials.
if [ -f "$HOME/.claude/.credentials.json" ] || [ -n "${ANTHROPIC_API_KEY:-}" ]; thenSuggested fix: Do not read credential stores or private files. If the task needs a credential, ask the user to provide it through the environment.
The text asks the agent to disclose its system prompt or hidden instructions.
Detection rate: catches naive exfiltration attempts that try to leak the system prompt.Suggested fix: Remove the request. A skill has no need to read or reveal the system prompt.
A script is downloaded and run in one step, so whatever the server returns at that moment is executed without review. Severity lowered from high because it appears in a code comment.
# - bun (install: curl -fsSL https://bun.sh/install | bash)Suggested fix: Download the script, check it (and its checksum or signature), then run the saved copy. Prefer a package manager.
The skill refers to a token store or the system keychain, a location that holds private data or credentials. Severity lowered from high because it appears in prose rather than code.
- Auth checks require the `claude` CLI plus either `~/.claude/.credentials.json` or `ANTHROPIC_API_KEY`.Suggested fix: Do not read credential stores or private files. If the task needs a credential, ask the user to provide it through the environment.
The cloud instance metadata service is contacted. On a cloud machine it hands out credentials. Severity lowered from high because it appears in prose rather than code.
- **IPv6 metadata bypass.** The URL constructor normalizes `::ffff:169.254.169.254` to `::ffff:a9fe:a9fe` (hex), which wasn't in the blocklist. Added both hex-encoded forms. From @mehmoodosman.Suggested fix: Remove access to the metadata service.
The skill refers to SSH keys, a location that holds private data or credentials. Severity lowered from high because it appears in prose rather than code.
- **Path traversal on /api/reload blocked.** Could previously read any file on disk (including ~/.ssh/id_rsa) by passing an arbitrary path in the JSON body. Now validates paths stay within cwd or tmpdir.Suggested fix: Do not read credential stores or private files. If the task needs a credential, ask the user to provide it through the environment.
The text uses language typical of jailbreak prompts.
- TypeScript/JS library for prompt injection, PII, jailbreak, profanity detectionSuggested fix: Remove the jailbreak framing.
The text tells the agent to ignore or override its existing instructions, a common prompt-injection pattern. Severity lowered from high because it appears in a code example.
- ignore (all |the )?(previous|above|prior) (instructions|rules|prompt)Suggested fix: Remove the instruction. A skill should add guidance for its task, not cancel the instructions the agent already follows.
The text tells the agent to disable safety checks, permission prompts or other safeguards. Severity lowered from high because it appears in a code example.
--dangerously-skip-permissions — bypass all permission checks (CI/automation)Suggested fix: Remove the instruction. If the task genuinely needs elevated permissions, say so plainly and let the user decide.
A text file was not stored in full (it is a symbolic link, exceeds the size limit or could not be fetched), so it was not fully scanned.
Suggested fix: Keep skill files below the size limit and avoid symbolic links.
- low
INT-MISSING-LICENSENeither the skill nor its repository states a license, so the terms of use are unclear.
Suggested fix: Add a LICENSE file to the repository or a `license` field to SKILL.md.
The text asks the agent to disclose its system prompt or hidden instructions. Severity lowered from high because the surrounding text warns against it.
4. **L5 canary token (`browse/src/security.ts`).** A random token injected into the system prompt at session start. Rolling-buffer detection across `text_delta` and `input_json_delta` streams catches the token if it shows up anywhere in Cl…Suggested fix: Remove the request. A skill has no need to read or reveal the system prompt.
The skill reads a .env file, which usually holds secrets. This is often legitimate, but check what is done with the values.
cp "$MAIN_WORKTREE/.env" "$REPO_ROOT/.env"Suggested fix: Do not read credential stores or private files. If the task needs a credential, ask the user to provide it through the environment.
A script is downloaded and run in one step, so whatever the server returns at that moment is executed without review. Severity lowered from high because the surrounding text warns against it.
Open your sidebar on Stack Overflow posts about prompt injection, read a Wikipedia article on SQL injection, browse a tutorial that walks through attack strings, the review banner stays quiet where before it fired. When a real hijack attem…Suggested fix: Download the script, check it (and its checksum or signature), then run the saved copy. Prefer a package manager.
The text asks the agent to disclose its system prompt or hidden instructions. Severity lowered from high because the surrounding text warns against it.
- `browse/src/security-classifier.ts` — Haiku model pinned to `claude-haiku-4-5-20251001` (no longer rolls forward silently via the `haiku` alias). `claude -p` now spawns from `os.tmpdir()` so CLAUDE.md project context doesn't leak into Ha…Suggested fix: Remove the request. A skill has no need to read or reveal the system prompt.
The text tells the agent to ignore or override its existing instructions, a common prompt-injection pattern. Severity lowered from high because it appears in a quotation.
- **Learnings prompt injection.** Three fixes: input validation (type/key/confidence allowlists), injection pattern detection in insight field (blocks "ignore previous instructions" etc.), and cross-project trust gate (only user-stated lea…Suggested fix: Remove the instruction. A skill should add guidance for its task, not cancel the instructions the agent already follows.
The skill reads a .env file, which usually holds secrets. This is often legitimate, but check what is done with the values.
cp .env.example .envSuggested fix: Do not read credential stores or private files. If the task needs a credential, ask the user to provide it through the environment.
The skill refers to a token store or the system keychain, a location that holds private data or credentials. Severity lowered from high because the surrounding text warns against it.
**Codex OAuth tokens DO NOT work for image generation.** Tested 2026-03-26: both the Images API and Responses API reject `~/.codex/auth.json` access_token with "Missing scopes: api.model.images.request". Codex CLI also has no native imageg…Suggested fix: Do not read credential stores or private files. If the task needs a credential, ask the user to provide it through the environment.
The text tells the agent to ignore or override its existing instructions, a common prompt-injection pattern. Severity lowered from high because the surrounding text warns against it.
| P18 | Haiku response contains prompt-injection (`"Ignore all prior instructions..."`) | Sanitize: only append lines that are substring matches of the original raw output |Suggested fix: Remove the instruction. A skill should add guidance for its task, not cancel the instructions the agent already follows.
The text tells the agent to ignore or override its existing instructions, a common prompt-injection pattern. Severity lowered from high because it appears in a quotation.
**Injection defense** (consistent with existing `bin/gstack-learnings-log` patterns): the `question_summary` and any free-form user feedback fields are sanitized against known prompt-injection patterns ("ignore previous instructions," "sys…Suggested fix: Remove the instruction. A skill should add guidance for its task, not cancel the instructions the agent already follows.
A link uses unencrypted http:// (bellard.org), so its content can be altered in transit.
* http://bellard.org/jslinux/Suggested fix: Use https://.
A link uses unencrypted http:// (www.apple.com), so its content can be altered in transit.
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">Suggested fix: Use https://.
The skill ships an image, font, media file or office document. It was not scanned.
Suggested fix: None required. Informational.
The skill ships an image, font, media file or office document. It was not scanned.
Suggested fix: None required. Informational.
The skill ships an image, font, media file or office document. It was not scanned.
Suggested fix: None required. Informational.
The skill ships an image, font, media file or office document. It was not scanned.
Suggested fix: None required. Informational.
The skill ships an image, font, media file or office document. It was not scanned.
Suggested fix: None required. Informational.
Some dependencies use version ranges, so what gets installed can change over time; ranges are not checked for vulnerabilities.
9 unpinned: @anthropic-ai/sdk, @huggingface/transformers, @ngrok/ngrok, diff, marked, playwright, puppeteer-core, xterm, xterm-addon-fitSuggested fix: Pin exact versions, or commit a lockfile.
SKILL.md pre-approves unrestricted shell access through `allowed-tools`, so the agent may run commands without asking.
allowed-tools: - Bash - Read - AskUserQuestionSuggested fix: Restrict the tool, for example `Bash(git status:*)`, or leave approval to the user.
Checks run
- prompt-injection
- dangerous-commands
- credentials
- network
- file-access
- integrity
- dependencies
Checks skipped
- gitleaks: gitleaks is not installed on this host.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 197k | Updated | Claude CodeCodex | Verified | Install |