š¦
Audit History
doc-sync-tool - 5 audits
Version comparison
Capability and finding changes across audited versions, newest first.
| Version | Date | Result | Review items | Change vs previous |
|---|---|---|---|---|
| v5 Latest | Jul 8, 2026, 12:44 PM | 2 confirmed | 8 | No capability change |
| v4 | Jul 8, 2026, 12:44 PM | 2 confirmed | 8 | Contains scripts |
| v3 | Jul 1, 2026, 03:31 AM | 1 confirmed | 4 | No capability change |
| v2 | Jul 1, 2026, 03:31 AM | 1 confirmed | 4 | Contains scripts |
| v1 | Feb 24, 2026, 09:43 AM | No confirmed findings | 1 | Baseline |
Audit version 5 Latest
Jul 8, 2026, 12:44 PM
Most static alerts are false positives from Markdown examples, JavaScript template literals, ANSI color codes, and hardcoded helper commands. Confirmed risks involve local file writes, watch-triggered synchronization, predictable test temporary paths, and unreviewed propagation of AI agent instruction files. No prompt injection attempt, network access, or credential exfiltration was found in the reviewed files.
13
Files scanned
1,747
Lines analyzed
12
Review items
0
False positives ignored
Confirmed security concerns (2)
High
Unreviewed Agent Instruction Overwrite
The sync logic selects the newest AGENTS.md, CLAUDE.md, or GEMINI.md file and writes its content to companion files without a diff, backup, or confirmation. This can propagate malicious or mistaken agent instructions across tools.
The code explicitly sorts by modification time and writes source content into other agent files. Watch mode automatically triggers the same sync path after file changes.
Medium
Broad Recursive Project Modification Scope
The tool scans from the current working directory or a wrapper-selected target directory, then recursively processes matching files. Running it from a broad parent directory can modify many projects unexpectedly.
CONFIG.scanPath is process.cwd(), scanDirectory recurses through child directories, and run-sync.sh changes into a caller-provided target before running sync.js.
Capability review items (8)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
Medium
Node.js fs operations Ā· 2 occurrences
fs.renameSync(file, newPath);
fs.renameSync renames every discovered Agents.md file under the current tree and can overwrite AGENTS.md without a prompt. This is an intended helper, but it creates real data-integrity risk.
Medium
Node.js fs operations Ā· 2 occurrences
fs.writeFileSync(targetFilePath, content, 'utf8');
writeFileSync overwrites an existing target agent file when content differs from the selected newest source. There is no diff, backup, or confirmation gate.
Medium
Synchronous file operations Ā· 2 occurrences
fs.writeFileSync(targetFilePath, content, 'utf8');
The synchronous write overwrites an existing target agent file without review. The risk is the write behavior, not the synchronous API choice.
Medium
Temp directory access
TEST_DIR="/tmp/doc-sync-test-$(date +%s)"
The test script uses a predictable /tmp path based on the current timestamp and later removes it recursively. This is test-only, but predictable temp paths are unsafe on shared systems.
Medium
File system watching
const watcher = fs.watch(dir, { recursive: true }, (eventType, filename) => {
fs.watch monitors the project tree and triggers synchronization on matching file changes. Persistent automatic sync increases the chance of unreviewed overwrites.
Risk Factors
āļø External commands (69)
EXAMPLES.md:23-24 EXAMPLES.md:24-27 EXAMPLES.md:27-28 EXAMPLES.md:28-34 EXAMPLES.md:392-394 EXAMPLES.md:254 README.md:230 rename-quick.js:5 rename-quick.js:14 rename-quick.js:33 rename-quick.js:25 rename-quick.js:44 rename-quick.js:51 rename-to-uppercase.sh:14 rename-to-uppercase.sh:22 rename-to-uppercase.sh:1 run-sync.sh:6 run-sync.sh:7 run-sync.sh:1 run-watch.sh:6 run-watch.sh:7 run-watch.sh:1 SKILL.md:10 SKILL.md:29-32 SKILL.md:32-36 SKILL.md:36-42 SKILL.md:42-46 SKILL.md:46-52 SKILL.md:52-56 SKILL.md:56-65 SKILL.md:65-69 SKILL.md:69 SKILL.md:78-80 SKILL.md:80-86 SKILL.md:86-90 SKILL.md:90 SKILL.md:98-100 SKILL.md:100-106 SKILL.md:106-109 sync.js:25 sync.js:60 sync.js:69 sync.js:75 sync.js:94 sync.js:97 sync.js:102 sync.js:108 sync.js:110 sync.js:115 sync.js:123 sync.js:124 sync.js:136 sync.js:147 sync.js:148 sync.js:149 sync.js:150 test.sh:8 test.sh:9 test.sh:76 test.sh:78 test.sh:1 watch.js:21 watch.js:50 watch.js:78 watch.js:90 watch.js:101 watch.js:111 watch.js:112 watch.js:113
š Filesystem access (31)
EXAMPLES.md:121 EXAMPLES.md:122 EXAMPLES.md:154 QUICKSTART.md:82 QUICKSTART.md:83 QUICKSTART.md:84 QUICKSTART.md:85 rename-quick.js:24 rename-quick.js:43 rename-quick.js:7 rename-quick.js:14 rename-quick.js:33 rename-to-uppercase.sh:18 rename-to-uppercase.sh:26 sync.js:31 sync.js:79 sync.js:90 sync.js:93 sync.js:101 sync.js:53 sync.js:79 sync.js:90 sync.js:93 sync.js:101 test.sh:26 test.sh:36 test.sh:48 test.sh:64 test.sh:77 test.sh:9 watch.js:63
Audited by: codex
Jul 8, 2026, 12:44 PM
Most static alerts are false positives from Markdown examples, JavaScript template literals, ANSI color codes, and hardcoded helper commands. Confirmed risks involve local file writes, watch-triggered synchronization, predictable test temporary paths, and unreviewed propagation of AI agent instruction files. No prompt injection attempt, network access, or credential exfiltration was found in the reviewed files.
13
Files scanned
1,747
Lines analyzed
12
Review items
0
False positives ignored
Confirmed security concerns (2)
High
Unreviewed Agent Instruction Overwrite
The sync logic selects the newest AGENTS.md, CLAUDE.md, or GEMINI.md file and writes its content to companion files without a diff, backup, or confirmation. This can propagate malicious or mistaken agent instructions across tools.
The code explicitly sorts by modification time and writes source content into other agent files. Watch mode automatically triggers the same sync path after file changes.
Medium
Broad Recursive Project Modification Scope
The tool scans from the current working directory or a wrapper-selected target directory, then recursively processes matching files. Running it from a broad parent directory can modify many projects unexpectedly.
CONFIG.scanPath is process.cwd(), scanDirectory recurses through child directories, and run-sync.sh changes into a caller-provided target before running sync.js.
Capability review items (8)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
Medium
Node.js fs operations Ā· 2 occurrences
fs.renameSync(file, newPath);
fs.renameSync renames every discovered Agents.md file under the current tree and can overwrite AGENTS.md without a prompt. This is an intended helper, but it creates real data-integrity risk.
Medium
Node.js fs operations Ā· 2 occurrences
fs.writeFileSync(targetFilePath, content, 'utf8');
writeFileSync overwrites an existing target agent file when content differs from the selected newest source. There is no diff, backup, or confirmation gate.
Medium
Synchronous file operations Ā· 2 occurrences
fs.writeFileSync(targetFilePath, content, 'utf8');
The synchronous write overwrites an existing target agent file without review. The risk is the write behavior, not the synchronous API choice.
Medium
Temp directory access
TEST_DIR="/tmp/doc-sync-test-$(date +%s)"
The test script uses a predictable /tmp path based on the current timestamp and later removes it recursively. This is test-only, but predictable temp paths are unsafe on shared systems.
Medium
File system watching
const watcher = fs.watch(dir, { recursive: true }, (eventType, filename) => {
fs.watch monitors the project tree and triggers synchronization on matching file changes. Persistent automatic sync increases the chance of unreviewed overwrites.
Risk Factors
āļø External commands (69)
EXAMPLES.md:23-24 EXAMPLES.md:24-27 EXAMPLES.md:27-28 EXAMPLES.md:28-34 EXAMPLES.md:392-394 EXAMPLES.md:254 README.md:230 rename-quick.js:5 rename-quick.js:14 rename-quick.js:33 rename-quick.js:25 rename-quick.js:44 rename-quick.js:51 rename-to-uppercase.sh:14 rename-to-uppercase.sh:22 rename-to-uppercase.sh:1 run-sync.sh:6 run-sync.sh:7 run-sync.sh:1 run-watch.sh:6 run-watch.sh:7 run-watch.sh:1 SKILL.md:10 SKILL.md:29-32 SKILL.md:32-36 SKILL.md:36-42 SKILL.md:42-46 SKILL.md:46-52 SKILL.md:52-56 SKILL.md:56-65 SKILL.md:65-69 SKILL.md:69 SKILL.md:78-80 SKILL.md:80-86 SKILL.md:86-90 SKILL.md:90 SKILL.md:98-100 SKILL.md:100-106 SKILL.md:106-109 sync.js:25 sync.js:60 sync.js:69 sync.js:75 sync.js:94 sync.js:97 sync.js:102 sync.js:108 sync.js:110 sync.js:115 sync.js:123 sync.js:124 sync.js:136 sync.js:147 sync.js:148 sync.js:149 sync.js:150 test.sh:8 test.sh:9 test.sh:76 test.sh:78 test.sh:1 watch.js:21 watch.js:50 watch.js:78 watch.js:90 watch.js:101 watch.js:111 watch.js:112 watch.js:113
š Filesystem access (31)
EXAMPLES.md:121 EXAMPLES.md:122 EXAMPLES.md:154 QUICKSTART.md:82 QUICKSTART.md:83 QUICKSTART.md:84 QUICKSTART.md:85 rename-quick.js:24 rename-quick.js:43 rename-quick.js:7 rename-quick.js:14 rename-quick.js:33 rename-to-uppercase.sh:18 rename-to-uppercase.sh:26 sync.js:31 sync.js:79 sync.js:90 sync.js:93 sync.js:101 sync.js:53 sync.js:79 sync.js:90 sync.js:93 sync.js:101 test.sh:26 test.sh:36 test.sh:48 test.sh:64 test.sh:77 test.sh:9 watch.js:63
Audited by: codex
Jul 1, 2026, 03:31 AM
Static analysis found many command, filesystem, network, obfuscation, and weak-crypto patterns. Review confirmed the runtime tool is a local documentation synchronizer, with no evidence of credential access, network exfiltration, prompt injection, or confirmed malicious intent. The real risk is broad recursive local file scanning and overwriting of agent instruction files, so publication should include a clear filesystem warning.
13
Files scanned
1,747
Lines analyzed
8
Review items
1
False positives ignored
Confirmed security concerns (1)
Low
Network Findings Are Reference Links Only
The only confirmed URL findings are documentation links to Node.js, PM2, and Git hook resources. No runtime code was found that sends data over the network.
Search found URLs only in markdown reference links. No fetch, HTTP client, webhook, or network transmission behavior was found in the executable files.
Capability review items (4)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
Medium
Recursive Documentation Overwrite Behavior
The sync script scans from the current working directory and copies the newest AGENTS.md, CLAUDE.md, or GEMINI.md content to peer files in each matched directory. This is expected functionality, but it can overwrite important agent instructions across a large tree if run from the wrong directory.
The code directly uses process.cwd() or a wrapper-supplied directory and writes synchronized content with fs.writeFileSync. The behavior is documented and not malicious, but the overwrite impact is clear.
Medium
Persistent File Watcher Can Propagate Unintended Changes
The watch mode recursively monitors the configured root and triggers synchronization when target files change. A mistaken edit or unexpected watcher event can propagate content to related agent instruction files automatically.
The watcher delegates to the same sync routine after filesystem events. The risk depends on where users start the watcher and whether they expect automatic writes.
Medium
Helper Rename Scripts Modify Files Broadly
The rename helper scripts search from the current directory and rename matching documentation files. They use fixed commands and quoted paths, but they can still rename files across an unexpectedly broad workspace.
The commands are hardcoded find operations, so command injection is unlikely. The confirmed concern is unintended local file renaming when launched from the wrong directory.
Low
Obfuscation And Weak Crypto Hits Appear Benign
Hex escape findings point to ANSI color escape sequences, and weak-crypto hits appear to match normal words such as sync in descriptions and code. No cryptographic routine or encoded payload was found.
The suspicious hex values are terminal color codes, and package metadata describes a sync tool. There is no semantic evidence of encryption, evasion, or packed code.
Static false positives ignored (1)
These static matches were dismissed by semantic review or matched schema-only tokens, so they are shown for transparency but do not drive the quality score.
Low
Documentation Command Examples Are Mostly False Positives
Most static command findings occur in markdown usage examples and shell snippets rather than hidden runtime behavior. They explain how to run Node, PM2, chmod, and Git hook commands for this local tool.
The cited lines are prose and fenced command examples. They do not execute by themselves, although users should still review commands before running them.
Risk Factors
ā” Contains scripts (4)
š Filesystem access (4)
āļø External commands (4)
Detected Patterns
Synchronous Filesystem Reads And WritesChild Process Use For Local Find CommandsRecursive Watch Mode
Audited by: codex
Jul 1, 2026, 03:31 AM
Static analysis found many command, filesystem, network, obfuscation, and weak-crypto patterns. Review confirmed the runtime tool is a local documentation synchronizer, with no evidence of credential access, network exfiltration, prompt injection, or confirmed malicious intent. The real risk is broad recursive local file scanning and overwriting of agent instruction files, so publication should include a clear filesystem warning.
13
Files scanned
1,747
Lines analyzed
8
Review items
1
False positives ignored
Confirmed security concerns (1)
Low
Network Findings Are Reference Links Only
The only confirmed URL findings are documentation links to Node.js, PM2, and Git hook resources. No runtime code was found that sends data over the network.
Search found URLs only in markdown reference links. No fetch, HTTP client, webhook, or network transmission behavior was found in the executable files.
Capability review items (4)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
Medium
Recursive Documentation Overwrite Behavior
The sync script scans from the current working directory and copies the newest AGENTS.md, CLAUDE.md, or GEMINI.md content to peer files in each matched directory. This is expected functionality, but it can overwrite important agent instructions across a large tree if run from the wrong directory.
The code directly uses process.cwd() or a wrapper-supplied directory and writes synchronized content with fs.writeFileSync. The behavior is documented and not malicious, but the overwrite impact is clear.
Medium
Persistent File Watcher Can Propagate Unintended Changes
The watch mode recursively monitors the configured root and triggers synchronization when target files change. A mistaken edit or unexpected watcher event can propagate content to related agent instruction files automatically.
The watcher delegates to the same sync routine after filesystem events. The risk depends on where users start the watcher and whether they expect automatic writes.
Medium
Helper Rename Scripts Modify Files Broadly
The rename helper scripts search from the current directory and rename matching documentation files. They use fixed commands and quoted paths, but they can still rename files across an unexpectedly broad workspace.
The commands are hardcoded find operations, so command injection is unlikely. The confirmed concern is unintended local file renaming when launched from the wrong directory.
Low
Obfuscation And Weak Crypto Hits Appear Benign
Hex escape findings point to ANSI color escape sequences, and weak-crypto hits appear to match normal words such as sync in descriptions and code. No cryptographic routine or encoded payload was found.
The suspicious hex values are terminal color codes, and package metadata describes a sync tool. There is no semantic evidence of encryption, evasion, or packed code.
Static false positives ignored (1)
These static matches were dismissed by semantic review or matched schema-only tokens, so they are shown for transparency but do not drive the quality score.
Low
Documentation Command Examples Are Mostly False Positives
Most static command findings occur in markdown usage examples and shell snippets rather than hidden runtime behavior. They explain how to run Node, PM2, chmod, and Git hook commands for this local tool.
The cited lines are prose and fenced command examples. They do not execute by themselves, although users should still review commands before running them.
Risk Factors
ā” Contains scripts (4)
š Filesystem access (4)
āļø External commands (4)
Detected Patterns
Synchronous Filesystem Reads And WritesChild Process Use For Local Find CommandsRecursive Watch Mode
Audited by: codex
Feb 24, 2026, 09:43 AM
Static analysis flagged 254 external_commands patterns and 31 filesystem operations, but evaluation confirms these are false positives. The 'Ruby/shell backtick execution' findings are markdown documentation examples in EXAMPLES.md, INDEX.md, README.md, QUICKSTART.md, and SKILL.md - not executable code. The 'hex-encoded characters' and 'weak cryptographic algorithm' findings are ANSI terminal color codes (e.g., \x1b[32m for green). The tool performs legitimate file synchronization using Node.js fs module with proper path handling via path.join() and path.dirname(). One minor concern: rename-quick.js uses child_process.execSync with hardcoded 'find' commands, which is acceptable but should be noted.
14
Files scanned
1,754
Lines analyzed
3
Review items
0
False positives ignored
Capability review items (1)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
Low
Shell Command Execution in Helper Script
The rename-quick.js helper script uses child_process.execSync to run 'find' commands for file discovery. While the arguments are hardcoded and not user-controllable, this represents a minor security consideration for a helper utility.
Risk Factors
āļø External commands (5)
š Filesystem access (5)
Audited by: claude