Security audit
Security checks for vulnerabilities and agentic risk
This skill needs review because conflicting payment instructions and a confirmed preview bug can cause unexpected credit spending.
Prefer a corrected package with consistent spending controls and an enforced return-risk scope. If proceeding, set ARI to ask before every charge, use the specialized quote-and-confirm workflow, choose manual collection for one-time requests, and review recurring schedules. Do not assume deepdive is a free preview.
scripts/ari.pyDeepdive's quote-only default can silently become paid analysis
Account credits can be consumed when the caller expects free aggregation and an analysis quote. The same helper also defeats the target-is-absent downgrade, and scripts/ari.py:1584 can then incorrectly state that no credits were deducted.
The substantive issue is conflicting authorization instructions, not merely the presence of additional commands: generic workflows permit automatic paid execution while the fixed return-risk entry requires a quote and explicit confirmation. These competing paths can cause an agent to spend credits outside the specialized workflow, although several monitoring commands have strong consent safeguards and some claimed code behaviors cannot be verified from this file.
Recurring monitoring and watch lifecycle management introduce persistent operational behavior and potential costs not disclosed by the evidence-analysis description. These are substantive additional capabilities rather than merely implementation details of reviewing product feedback. The chunk contains command examples, so it does not establish that watches were actually created or charges incurred, nor does it reveal the underlying data collected. It nevertheless exposes an undeclared monitoring workflow.
The declared purpose describes evidence-based analysis and recommendations, but does not disclose recurring monitoring or management of priced refresh jobs. These are material operational capabilities rather than merely internal analysis details. The snippet does not establish that charges actually occur or reveal what each refresh analyzes, but it explicitly exposes scheduled-watch lifecycle operations and price acceptance.
声明描述的是评论证据分析，而代码片段提供了修改持久化采集计划的操作，包括开启每周自动采集。这超出了单次分析所需的普通实现细节，增加了未披露的后台自动化及潜在持续费用。片段未展示底层实现，因此无法确认任务实际如何执行，但其明确提供的调度能力已构成描述不匹配。
The declared purpose is review-evidence analysis, but the snippet exposes watch-management and pricing functionality not disclosed in that description. The supplied code does not show how those capabilities support the declared analysis. Because the implementation of scripts/ari.py is absent, this assessment is limited to the visible commands and their accompanying comment; it does not establish that watches are created or charges incurred.
The declared purpose describes review-evidence analysis, but does not disclose watch-related inspection or cost estimation. The supplied commands expose those additional capabilities without demonstrating a supporting role in the declared analysis. The snippet does not show watch creation, billing transactions, or inventory, refund, or customer-service execution.
Scheduled watch creation and lifecycle changes introduce persistent operational behavior and potential recurring costs that the description does not disclose. These are materially broader than merely analyzing review evidence and suggesting improvements. Digest, event, and series retrieval could support the declared analysis, but the excerpt does not show their contents. Because only CLI invocations are supplied, the underlying implementation and whether charges actually occur cannot be verified.
The comparison workflow presents competitor binding as a step toward comparison but states that binding starts weekly automatic collection without explicitly requiring separate recurring-charge consent at that step. Approval of a one-time comparison should not implicitly authorize an ongoing collection relationship; general scheduling safeguards mitigate but do not resolve this ambiguity.
run_analysis changes confirm from false to true when the server returns a truthy autoConfirm value and sufficient balance, allowing paid analysis without --confirm. This contradicts the CLI's explicit-confirmation contract and also affects deepdive, whose supposed quote-only fallback can consequently execute a charged analysis.
In context, this entry preserves a policy that allows paid operations without confirmation based on a server-provided autoConfirm flag, without establishing prior user authorization for that spending. This can consume account credits unexpectedly, although the changelog alone does not establish whether this specialized package exposes the affected operations.
- 重新发布：各市场首批上架的 1.5.0 包早于免确认修复，新注册的免费版用户仍会被要求先确认报价。
1.5.1 与当前官网下发的修复版一致，市场侧以新版本号覆盖旧包。
- 免确认只认服务端报价返回的 `autoConfirm`；采集报价扣除首次采集免费页；
非美国站首次采集可用赠送积点时不再提示「赠送积点不可用」（行为与修复后的 1.5.0 相同）。
- 命令、字段与计费规则不变，不提高最低支持版本。
The documented policy lets an initial report request trigger collection, analysis, and credit deductions without a price confirmation, and explicitly removes an amount cap for the first N paid generations. A sufficient balance and a post-operation charges disclosure do not establish informed authorization to spend; gifted credits mitigate immediate financial exposure but remain consumable account resources.
- 行为变更：新用户（注册并验证邮箱后）第一次要 VOC，只给 ASIN 即可（站点默认美国站），
`voc` 直接执行——自动补采集（默认 3 页）、生成报告、按服务端定价扣费并归档，
不再停在报价等确认。费用在报告返回的 `charges` 字段透明展示（采集/分析/合计/余额前后）。
- 免确认与否**只认服务端报价返回的 `autoConfirm` 布尔**：服务端策略前 N 次付费生成不看金额上限
（新用户赠送积点按设计覆盖「3 页采集 + 一份 VOC」）；用户自设阈值优先、显式要求「每次先问」
时仍先报价；余额不足不自动。客户端不再从 `autoConfirmNote` 文案或 `autoConfirmRemaining`
自行推断——文案匹配会把 Free 小额提示里的「先确认」误判成用户要求每次询问，恰好挡住新用户。
This match is part of the same documented automatic-spending policy, rather than an independent vulnerability. Honoring user-set thresholds and explicit requests for confirmation reduces risk, but users who have not configured either can still incur deductions under a default server policy without approving the cost.
不再停在报价等确认。费用在报告返回的 `charges` 字段透明展示（采集/分析/合计/余额前后）。
- 免确认与否**只认服务端报价返回的 `autoConfirm` 布尔**：服务端策略前 N 次付费生成不看金额上限
（新用户赠送积点按设计覆盖「3 页采集 + 一份 VOC」）；用户自设阈值优先、显式要求「每次先问」
时仍先报价；余额不足不自动。客户端不再从 `autoConfirmNote` 文案或 `autoConfirmRemaining`
自行推断——文案匹配会把 Free 小额提示里的「先确认」误判成用户要求每次询问，恰好挡住新用户。
- 采集报价同口径扣除服务端 `trialFreePages`（首次采集试用免费页），避免新用户被误报「积点不足」。
- 非美国站：服务端 `credits.gift` 策略下用户首次采集不限站点，`credits/balance` 返回 `giftAllSites`，
This is a duplicate match on the same automatic-spending policy, not an additional attack path. Using a structured boolean rather than interpreting explanatory text is sound, but that boolean does not establish user consent when it can be enabled by a default server policy.
不再停在报价等确认。费用在报告返回的 `charges` 字段透明展示（采集/分析/合计/余额前后）。
- 免确认与否**只认服务端报价返回的 `autoConfirm` 布尔**：服务端策略前 N 次付费生成不看金额上限
（新用户赠送积点按设计覆盖「3 页采集 + 一份 VOC」）；用户自设阈值优先、显式要求「每次先问」
时仍先报价；余额不足不自动。客户端不再从 `autoConfirmNote` 文案或 `autoConfirmRemaining`
自行推断——文案匹配会把 Free 小额提示里的「先确认」误判成用户要求每次询问，恰好挡住新用户。
- 采集报价同口径扣除服务端 `trialFreePages`（首次采集试用免费页），避免新用户被误报「积点不足」。
- 非美国站：服务端 `credits.gift` 策略下用户首次采集不限站点，`credits/balance` 返回 `giftAllSites`，
This entry extends the same confirmation bypass to multiple analysis types, potentially broadening unexpected credit consumption beyond the initial VOC workflow. The risk concerns authorization of paid actions, not autonomous analysis itself; actual exposure in this specialized package depends on its command restrictions.
为 true 时赠送积点计入 `nonUsCollectCredits`、`siteNote` 为空；UK/DE/JP 卖家的第一份 VOC 同样一步直出。
- 余额不足仍在执行前拦截（402）并引导充值；定价以服务端配置为准，代码常量仅为兜底。
- 一致性：`analyze` 各类型（insight / keywords / trend / variant / compare）走同一口径，
`autoConfirm=true` 即直接生成并返回 `charges`。
- 文档：移除所有写死的单价与合计示例。采集/分析单价与页数由后台 `pricing` 配置决定，
一律以本次报价返回的 `credits` / `totalCredits` 为准，不得凭文档或记忆报数。
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.
- 能力层收窄：44 个专属包不再只是「声明窄」，包体 `scripts/ari.py` 同时按各自契约裁剪。
构建期用 AST 删除文档未承诺的命令、它们的解析器注册块，以及只被它们引用的辅助函数与
常量；专属包因此不再包含 `export`（任意路径写文件）、`autoconfirm`（撤销逐次确认）、
`workbench`（改写差评处理状态）、`alerts`（标记预警已读）、`competitors`、`schedule`、
`voc` / `analyze` / `charts` / `deepdive` 等母版命令。
- 新增运行期兜底闸：专属包在包内写入 `allowedCommands` 白名单，`ari.py`
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.
def cmd_autoconfirm(args):
"""免确认阈值：不带参数=查看；`autoconfirm 50`=50 积点以内不问；`autoconfirm off`=每次都问；`autoconfirm default`=恢复默认。"""
value = (args.value or "").strip().lower()
if value:
if value in ("off", "ask", "0"):
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.
def cmd_autoconfirm(args):
"""免确认阈值：不带参数=查看；`autoconfirm 50`=50 积点以内不问；`autoconfirm off`=每次都问；`autoconfirm default`=恢复默认。"""
value = (args.value or "").strip().lower()
if value:
if value in ("off", "ask", "0"):
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.
def cmd_autoconfirm(args):
"""免确认阈值：不带参数=查看；`autoconfirm 50`=50 积点以内不问；`autoconfirm off`=每次都问；`autoconfirm default`=恢复默认。"""
value = (args.value or "").strip().lower()
if value:
if value in ("off", "ask", "0"):
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.
limit = int(value)
except ValueError:
emit(error_obj("ARI_BAD_ARGUMENT", 0, "参数不对",
"用法：autoconfirm 50（50 积点以内不问）/ autoconfirm off（每次都问）/ autoconfirm default（恢复默认）"),
args.compact)
return
out = request_json("PUT", "/api/v1/user/autoconfirm", {"limit": limit})
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.
limit = int(value)
except ValueError:
emit(error_obj("ARI_BAD_ARGUMENT", 0, "参数不对",
"用法：autoconfirm 50（50 积点以内不问）/ autoconfirm off（每次都问）/ autoconfirm default（恢复默认）"),
args.compact)
return
out = request_json("PUT", "/api/v1/user/autoconfirm", {"limit": limit})
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.
limit = int(value)
except ValueError:
emit(error_obj("ARI_BAD_ARGUMENT", 0, "参数不对",
"用法：autoconfirm 50（50 积点以内不问）/ autoconfirm off（每次都问）/ autoconfirm default（恢复默认）"),
args.compact)
return
out = request_json("PUT", "/api/v1/user/autoconfirm", {"limit": limit})
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.
return
out = request_json("PUT", "/api/v1/user/autoconfirm", {"limit": limit})
else:
out = request_json("GET", "/api/v1/user/autoconfirm")
if ok(out) and isinstance(data_of(out), dict):
d = data_of(out)
d["note"] = AUTOCONFIRM_MODE_NOTE.get(d.get("mode"), AUTOCONFIRM_MODE_NOTE["ask"]).format(
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.
p = sub.add_parser("check", parents=[common], help="验证 Key、账户和积点余额")
p.set_defaults(fn=cmd_check)
p = sub.add_parser("autoconfirm", parents=[common],
help="免确认阈值：autoconfirm 50 / autoconfirm off / autoconfirm default；不带参数查看")
p.add_argument("value", nargs="?", default="")
p.set_defaults(fn=cmd_autoconfirm)
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.
p.set_defaults(fn=cmd_check)
p = sub.add_parser("autoconfirm", parents=[common],
help="免确认阈值：autoconfirm 50 / autoconfirm off / autoconfirm default；不带参数查看")
p.add_argument("value", nargs="?", default="")
p.set_defaults(fn=cmd_autoconfirm)
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.
p.set_defaults(fn=cmd_check)
p = sub.add_parser("autoconfirm", parents=[common],
help="免确认阈值：autoconfirm 50 / autoconfirm off / autoconfirm default；不带参数查看")
p.add_argument("value", nargs="?", default="")
p.set_defaults(fn=cmd_autoconfirm)