Back to skill

Security audit

Session Wrap-Up Premium

Security checks for vulnerabilities and agentic risk

Overview

The skill is purpose-aligned as a session wrap-up tool, but it needs Review because it can publish the whole repository and contains unsafe command handling that could run unintended shell commands.

Install only if you are comfortable with a skill that writes persistent notes and can stage, commit, and push the entire current Git repository. Treat the current version as needing fixes before routine use: disable or remove automatic push, replace shell=True Git calls with argument arrays, honor push/update controls, and review staged files before publishing.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (5)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/run.py:84
Finding

OS Command Injection Through the Git Commit Message

Content
View full analysis
/tmp/skill-command-output; # ``` produces a command equivalent to: ```sh git commit -m "summary"; id > /tmp/skill-command-output; #" ``` The injected `id` command is therefore executed independently of Git. ### Attack Path 1. An attacker or untrusted caller invokes t ...[truncated 918 chars]
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Error
Location
scripts/run.py:155
Finding

The Explicit Push Opt-Out Is Ignored

Content
View full analysis
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Error
Location
scripts/run.py:80
Finding

Indiscriminate Git Staging Can Publish Unrelated Sensitive Files

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/run.py:148
Finding

PARA Update Controls and Target Selection Are Not Enforced

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/run.py:161
Finding

Undefined Variable Crashes Wrap-Up After Irreversible Side Effects

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Rogue AgentSelf-Modification, Session Persistence
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
Findings (11)

Context-Inappropriate Capability

Critical
Category
Not specified by scanner
Confidence
100% confidence
Finding

The commit message comes from input_data and is directly interpolated into git commit -m "{message}" before being executed by a shell. An attacker can inject quotes and shell operators to run arbitrary commands, making this a direct command injection vulnerability with repository modification and potential network-assisted exfiltration through subsequent push.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The skill’s described wrap-up behavior includes git add -A, commit, and push across the whole workspace, which can capture and exfiltrate unrelated or sensitive files well beyond a narrow session summary. The mismatch between friendly productivity framing and broad repository mutation makes the operation more dangerous because users may not realize the full extent of side effects before invoking it.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill defines a generic shell execution primitive for git operations even though the stated purpose is session wrap-up. That broadens capability unnecessarily: once combined with user-influenced inputs elsewhere in the script, the skill can execute shell-interpreted commands outside normal wrap-up behavior.

Content

No source excerpt is available for this finding.

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
97% confidence
Finding

This is a tool-parameter abuse issue because the subprocess invocation accepts a free-form shell command string instead of constrained parameters. In context, the skill is supposed to do structured session wrap-up tasks, but this design permits arbitrary shell interpretation and becomes especially dangerous when fed the user-controlled commit message path.

Content

Scanner excerpt · scripts/run.py (reported line 22)May include surrounding context.

python
OPEN_LOOPS = NOTES_DIR / "open-loops.md"

def run(cmd):
    res = subprocess.run(cmd, shell=True, capture_output=True, text=True)
    return res.returncode, res.stdout, res.stderr

def ensure_dirs():

Undeclared Tool Scope

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.

Session Persistence

Medium
Category
Rogue Agent
Confidence
83% confidence
Finding

This skill is intentionally designed for session persistence, but that persistence becomes risky because it stores durable logs, long-term memory, and PARA notes that may include sensitive operational context. In a wrap-up skill, retaining context is expected, yet without minimization and redaction controls it can accumulate confidential data and increase exposure if the repo is shared or pushed remotely.

Content

Scanner excerpt · SKILL.md (reported line 53)May include surrounding context.

md
enum: ["append", "replace"]
          default: "append"
      required: ["section", "content"]
    permission: workspace_write
---

# Session Wrap-Up Premium

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

Automatic file modification plus git push without an explicit privacy warning creates a real risk of unintentionally publishing secrets, internal notes, or sensitive session data to a remote repository. In the context of a memory/logging skill, the danger is elevated because the content being persisted may include summaries of prior work, credentials, paths, or incident details.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
96% confidence
Finding

The explicit design choice of performing commit and push 'automatic, no confirmation' enables autonomous externalization of workspace state without a final user checkpoint. That is dangerous because a single invocation can irreversibly publish data or alter repository history even if the generated summary or selected files are wrong.

Content

Scanner excerpt · SKILL.md (reported line 139)May include surrounding context.

git push

text

**Automatic**, no confirmation (by design — you already approved by running wrap_up).

### 5. Report Summary

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
98% confidence
Finding

The helper executes shell commands with shell=True, which means any metacharacters in command strings are interpreted by the shell. In this file, that becomes dangerous because later code interpolates a user-influenced commit message into a git command, turning this generic subprocess pattern into a command injection path with local code execution and possible data exfiltration via git push.

Content

Scanner excerpt · scripts/run.py (reported line 22)May include surrounding context.

python
OPEN_LOOPS = NOTES_DIR / "open-loops.md"

def run(cmd):
    res = subprocess.run(cmd, shell=True, capture_output=True, text=True)
    return res.returncode, res.stdout, res.stderr

def ensure_dirs():

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The function performs git add, commit, and push automatically, causing repository state changes and outbound network activity without any confirmation gate. In an agent skill context, this is riskier because wrap-up may be triggered routinely, so unintended commits or pushes can leak sensitive notes and memory files to remotes.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The script automatically writes summaries, memory highlights, and open loops to workspace files without notifying the user. While expected for a wrap-up tool, it can still create privacy and integrity issues by persisting sensitive or incorrect data into tracked files that may later be committed and pushed.

Content

No source excerpt is available for this finding.