Back to skill

Security audit

AI Clipping Skill

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill suspicious because scanners report weaknesses in disclosure checks, file cleanup, data handling, and tool scope; the scan was partial, so the evidence is incomplete.

Consider the reported disclosure, file-cleanup, and credential-handling weaknesses before installing. The scan covered 40 of 41 components; one file was only partially inspected, so the evidence is incomplete and does not establish that the skill is safe.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
Findings (36)

Harmful Content Injection

Critical
Category
Prompt Injection
Confidence
90% confidence
Finding

This content may contain harmful instructions that could cause physical harm if followed. CRITICAL: Review carefully before use.

Content

Scanner excerpt · examples/orbit-campaign/README.md (reported line 72)May include surrounding context.

md
Without the two proofreading marks, `make_clips.py` warns that the Groundhog Day clip ends
mid-sentence; that is the check doing its job. Remove the old `plan.md` and `plan.html` first, or
plan.py stops rather than overwrite them.

Clips you cut yourself get new fingerprints, which is fine: they are new files.

Tainted flow: 'req' from os.environ.get (line 169, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/transcribe.py (reported line 172)May include surrounding context.

python
req = urllib.request.Request(url, data=bytes(body), method="POST", headers={
        "Authorization": f"Bearer {key}", "Content-Type": f"multipart/form-data; boundary={boundary}"})
    try:
        with urllib.request.urlopen(req, timeout=600) as r:
            return json.loads(r.read())
    except urllib.error.HTTPError as e:
        die(f"transcription API error {e.code}: {e.read().decode(errors='replace')[:400]}")

Context Leakage

High
Category
Data Exfiltration
Confidence
85% confidence
Finding

Code or instructions that leak agent conversation context to external services, potentially exposing sensitive user interactions.

Content

Scanner excerpt · scripts/upload_parts.py (reported line 2)May include surrounding context.

python
#!/usr/bin/env python3
"""Upload one file to a WoopSocial media upload session (standard library only).

WoopSocial uploads work in three steps:
  1. The agent calls `media_uploads_create_session` with the projectId and the file size in bytes.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: env, file_read, file_write, mcp, network, shell.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

The transcription workflow lacks a warning that cloud transcription sends footage audio to a third party.

Medium
Category
Output Handling
Confidence
85% confidence
Finding

The workflow offers cloud transcription without explicitly explaining that footage audio leaves the user's device or requiring approval before that transfer. Providing an API key is not necessarily consent to upload particular recordings, which could include private or unreleased material; the later posting approval does not cover this earlier disclosure.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · examples/orbit-campaign/checklist.md (reported line 19)May include surrounding context.

md
10. **Engagement:** at least 3% of views (likes, comments and shares).
11. **Time live:** stays public and unchanged for 30 days; views count for 30 days from the day a clip crosses the payout threshold.
12. **Proof to keep:** analytics screenshot with views and audience countries at 72 hours. Screenshots at 48 hours, 72 hours and 7 days per post either way.
13. **Approval:** no approval step before posting.
14. **Sound:** no required sound.
15. **Banned:** giveaway, #fyp.
16. **Submitting:** within 30 minutes of posting; one post goes into one campaign, once.

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · references/woopsocial.md (reported line 38)May include surrounding context.

md
|---|---|
| Claude (claude.ai, Desktop, Cowork) | Add WoopSocial from the [Claude directory](https://claude.ai/directory/woopsocial), click Connect and sign in |
| Cursor | Install from [cursor.directory](https://cursor.directory/plugins/woopsocial-social-media-scheduler) and sign in |
| Claude Code | `claude mcp add --transport http woopsocial "https://api.woopsocial.com/mcp"`, then restart with `claude --continue` (keeps the conversation), `/mcp`, select woopsocial, Authenticate |
| Codex | `codex mcp add woopsocial --url https://api.woopsocial.com/mcp`, then `codex mcp login woopsocial`, then start a new session (`codex resume --last` keeps this one) |
| OpenClaw | Install the [WoopSocial skill on ClawHub](https://clawhub.ai/woopsocial/skills/social-media-scheduler), or add the MCP URL to the agent's MCP config |
| Agents without sign-in support | Create an API key at [app.woopsocial.com/api-access](https://app.woopsocial.com/api-access) and use `https://api.woopsocial.com/mcp?api_k=

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · references/woopsocial.md (reported line 39)May include surrounding context.

md
(claude.ai, Desktop, Cowork) | Add WoopSocial from the [Claude directory](https://claude.ai/directory/woopsocial), click Connect and sign in |
| Cursor | Install from [cursor.directory](https://cursor.directory/plugins/woopsocial-social-media-scheduler) and sign in |
| Claude Code | `claude mcp add --transport http woopsocial "https://api.woopsocial.com/mcp"`, then restart with `claude --continue` (keeps the conversation), `/mcp`, select woopsocial, Authenticate |
| Codex | `codex mcp add woopsocial --url https://api.woopsocial.com/mcp`, then `codex mcp login woopsocial`, then start a new session (`codex resume --last` keeps this one) |
| OpenClaw | Install the [WoopSocial skill on ClawHub](https://clawhub.ai/woopsocial/skills/social-media-scheduler), or add the MCP URL to the agent's MCP config |
| Agents without sign-in support | Create an API key at [app.woopsocial.com/api-access](https://app.woopsocial.com/api-access) and use `https://api.woopsocial.com/mcp?api_key=%2A%2A%2A |

External Transmission

Medium
Category
Data Exfiltration
Confidence
96% confidence
Finding

The fallback places a password-equivalent API key in the MCP URL query string. Such URLs can be persisted in agent configuration, diagnostic output, or request logs, exposing a credential that may authorize social-account operations; HTTPS does not prevent these local or server-side disclosures.

Content

Scanner excerpt · references/woopsocial.md (reported line 41)May include surrounding context.

md
| Claude Code | `claude mcp add --transport http woopsocial "https://api.woopsocial.com/mcp"`, then restart with `claude --continue` (keeps the conversation), `/mcp`, select woopsocial, Authenticate |
| Codex | `codex mcp add woopsocial --url https://api.woopsocial.com/mcp`, then `codex mcp login woopsocial`, then start a new session (`codex resume --last` keeps this one) |
| OpenClaw | Install the [WoopSocial skill on ClawHub](https://clawhub.ai/woopsocial/skills/social-media-scheduler), or add the MCP URL to the agent's MCP config |
| Agents without sign-in support | Create an API key at [app.woopsocial.com/api-access](https://app.woopsocial.com/api-access) and use `https://api.woopsocial.com/mcp?api_key=%2A%2A%2A |

The API key is a password: never paste it into shared chats or commit it. Claude Code and Codex
load new tools only in a new session, and a restart stops a running script: never add the

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
80% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · references/woopsocial.md (reported line 54)May include surrounding context.

md
**Connect every account in one pass.** `projects_list`, then `oauth_create_authorization` for each
missing platform, all links in one message. When they're done, `social_accounts_list` once and put
each `socialAccountId` in `clipjob.json`. Never ask the user to paste ids.

## 3. Read each TikTok account's limits

The supposedly example-only disclosure exemption is enabled entirely by editable campaign fields.

Medium
Category
Excessive Agency
Confidence
98% confidence
Finding

The example exemption is authorized solely by three editable campaign fields, rather than by trusted provenance. A crafted campaign can satisfy those checks and have “EXAMPLE ONLY: not a real ad” accepted instead of a genuine on-screen paid disclosure, undermining the claimed restriction even though other disclosure requirements remain.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · scripts/campaign.py (reported line 456)May include surrounding context.

python
items.append("**Proof to keep:** " + ("; ".join(c["proof"]) if c.get("proof") else "nothing extra named")
                 + ". Screenshots at 48 hours, 72 hours and 7 days per post either way.")
    items.append("**Approval:** the brand approves each clip before it is posted (the plan writes a packet to send)."
                 if c.get("approval_before_posting") else "**Approval:** no approval step before posting.")
    snd = c.get("required_sound")
    if snd:
        plats = ", ".join(LIMITS[p]["label"] for p in snd["platforms"] if p in LIMITS)

The documented footage-overlap limit is not enforced for cross-platform edits with different hooks and captions.

Medium
Category
Excessive Agency
Confidence
98% confidence
Finding

When footage overlap exceeds the campaign's max_overlap, cross-platform posts with different hooks and caption bodies receive only a warning rather than an error. Because warnings do not block posting, edits can pass validation despite violating the campaign's overlap limit, potentially causing rejected payouts or account penalties.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/check_setup.py (reported line 27)May include surrounding context.

python
else:
        line("ok", "Python", platform.python_version())
    if shutil.which("ffmpeg") and shutil.which("ffprobe"):
        out = subprocess.run(["ffmpeg", "-version"], capture_output=True, text=True).stdout.split("\n")[0]
        line("ok", "ffmpeg", (out.split() + ["", "", "?"])[2] + ": reads, cuts and checks the clips")
        if not cliplib.ffmpeg_has("encoders", "libx264"):
            line("note", "libx264", "this ffmpeg can check clips but cutting new ones needs libx264. Mac: brew "

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
70% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · scripts/check_setup.py (reported line 34)May include surrounding context.

python
"install ffmpeg. Windows: winget install Gyan.FFmpeg. Linux: your distro's full ffmpeg package")
    else:
        line("missing", "ffmpeg", "Mac: brew install ffmpeg. Windows: winget install Gyan.FFmpeg (then open a "
             "new terminal). Linux: sudo apt install ffmpeg")
        ok = False
    try:
        import PIL

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
70% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · scripts/cliplib.py (reported line 257)May include surrounding context.

python
import shutil
    if not (shutil.which("ffmpeg") and shutil.which("ffprobe")):
        die("ffmpeg not found. Mac: brew install ffmpeg. Windows: winget install Gyan.FFmpeg (then open a "
            "new terminal). Linux: sudo apt install ffmpeg")


@functools.lru_cache(maxsize=None)

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
70% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · scripts/make_clips.py (reported line 102)May include surrounding context.

python
def check_ffmpeg():
    if not shutil.which("ffmpeg") or not shutil.which("ffprobe"):
        die("ffmpeg is not installed. Mac: brew install ffmpeg. Windows: winget install Gyan.FFmpeg. "
            "Linux: sudo apt install ffmpeg.")
    load_pillow()

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
70% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · scripts/transcribe.py (reported line 45)May include surrounding context.

python
def need_ffmpeg():
    if not shutil.which("ffmpeg") or not shutil.which("ffprobe"):
        die("ffmpeg is not installed. Mac: brew install ffmpeg. Windows: winget install ffmpeg. "
                 "Linux: sudo apt install ffmpeg (or your distro's package).")


def duration(path):

Tainted flow: 'tmp' from os.environ.get (line 186, credential/environment) → open (file write)

Medium
Category
Data Flow
Confidence
89% confidence
Finding

The environment-derived destination is not inherently unsafe, but opening the predictable path '.tmp' with mode 'w' follows existing symbolic links and truncates their targets. An attacker able to create entries in the destination directory could overwrite another file writable by the agent; concurrent writers can also interfere with the same temporary file.

Content

Scanner excerpt · scripts/cliplib.py (reported line 188)May include surrounding context.

python
os.makedirs(os.path.dirname(os.path.abspath(path)), exist_ok=True)
    tmp = path + ".tmp"
    try:
        with open(tmp, "w", encoding="utf-8") as fh:
            json.dump(data, fh, ensure_ascii=False, indent=1)
        os.replace(tmp, path)
    finally:

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/cliplib.py (reported line 263)May include surrounding context.

python
@functools.lru_cache(maxsize=None)
def ffmpeg_has(kind, name):
    """Whether this ffmpeg build has an encoder or filter."""
    r = subprocess.run(["ffmpeg", "-hide_banner", f"-{kind}"], capture_output=True, text=True,
                       encoding="utf-8", errors="replace")
    return bool(re.search(rf"\s{re.escape(name)}\s", r.stdout))

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/cliplib.py (reported line 306)May include surrounding context.

python
if not os.path.getsize(path):
        die(f"{name} is an empty file (0 bytes)")
    need_ffmpeg()
    r = subprocess.run(["ffprobe", "-v", "error", "-show_streams", "-show_format", "-of", "json", path],
                       capture_output=True, text=True, encoding="utf-8", errors="replace")
    try:
        d = json.loads(r.stdout or "{}")

The documented campaign-level rate fallback is not implemented, causing earnings estimates to use zero when only a campaign-wide rate is provided.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The terms() docstring at L760–L761 says platform-specific financial terms fall back to the campaign's values. Although minimum payout, cap, and view minimum use that fallback, L768 reads rate_per_1k only from the platform object and otherwise returns zero, contradicting the documented behavior and potentially understating expected earnings.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/make_clips.py (reported line 93)May include surrounding context.

python
def run(cmd, **kw):
    r = subprocess.run(cmd, capture_output=True, text=True, encoding="utf-8", errors="replace", **kw)
    if r.returncode != 0:
        die("ffmpeg failed: " + (r.stderr.strip().splitlines() or ["no message"])[-1][-600:])
    return r.stdout

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/make_clips.py (reported line 380)May include surrounding context.

python
urllib.request.urlretrieve(FONT_URL + FONT_FILES[0], path)
    except Exception:
        if shutil.which("curl"):
            subprocess.run(["curl", "-sfL", FONT_URL + FONT_FILES[0], "-o", path], check=False)
    if os.path.exists(path) and os.path.getsize(path) > 10000:
        return ImageFont.truetype(path, size)
    warn("caption font missing; using Pillow's default font")

Omitting the campaign file bypasses the documented mandatory paid-partnership disclosure.

Medium
Category
Excessive Agency
Confidence
98% confidence
Finding

When the campaign field is absent or empty, campaign_screen returns empty settings before enforcing the paid-partnership disclosure. An untrusted or incorrectly generated clip specification can therefore produce undisclosed promotional clips, undermining this skill's central compliance guarantee; any downstream posting checks are not demonstrated in this file.

Content

No source excerpt is available for this finding.