Pentest Api Attacker
ReviewAudited by ClawScan on May 10, 2026.
Overview
The skill has authorization safeguards, but the included script appears to generate generic high-severity API findings without performing the advertised API tests.
Install or run this only if you understand it appears to be a scaffold rather than a real API testing implementation. Use dry-run first, verify the missing shared dependency, provide only authorized in-scope targets, and do not treat its generated findings as verified vulnerabilities without independent evidence.
Findings (4)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Running live checks against APIs without authorization can create legal, operational, or service-impact risk.
The skill is explicitly dual-use and intended for live API security testing, but it also discloses scope and authorization controls.
Validate scope before any active action ... require explicit --i-have-authorization for live execution ... This skill executes real security testing tools against live targets.
Use only with written authorization, verify scope.json before running, and keep --dry-run enabled until the target and checks are confirmed.
Users or downstream agents could treat placeholder findings as real security results, causing false reports, wasted remediation effort, or misplaced trust in the assessment.
The script constructs a generic high-severity finding with placeholder evidence and writes it to artifacts; the visible code does not perform the advertised API testing before producing that finding.
'title':'Pentest API Attacker identified a security weakness' ... 'severity':'High' ... 'proof_of_concept':'python scripts/api_attacker.py --dry-run' ... write_placeholder_artifact(... 'findings':[finding])
Do not rely on generated findings as evidence of vulnerabilities until the skill implements real checks and clearly separates placeholders from verified results.
The skill may fail if the sibling dependency is absent, or it may execute unreviewed local shared code whose behavior is not visible in this review.
Core behavior, including scope validation, input loading, output path resolution, and artifact writing, depends on a shared module that is not included in the provided manifest.
SHARED_DIR=Path(__file__).resolve().parents[2]/"autonomous-pentester"/"shared" ... from pentest_common import load_payload,render_result,resolve_artifact_path,resolve_output_file,validate_scope,write_placeholder_artifact
Verify the provenance and contents of the autonomous-pentester shared module before running, or require the skill to bundle or pin the dependency.
Generated reports may retain sensitive API specifications, tokens, or project context if the selected input path contains them, especially if artifacts are later shared or consumed by other skills.
Whatever the helper loads from the input path is persisted into generated artifacts; the default input path is the current directory and the loading boundaries are not visible in the provided code.
p.add_argument('--input',default='.') ... payload=load_payload(args.input) ... 'input_payload':payloadPass a narrow, sanitized input file or directory, store outputs in a protected location, and inspect artifacts before sharing them downstream.