Amazon Competitor Spy

Amazon competitor intelligence agent. Tracks competitor ASINs over time — price changes, BSR movements, review count velocity, listing changes, new images/va...

Security Scan

The skill's description and data-storage behavior are plausible for an ASIN-tracking tool, but the SKILL.md references a 'spy' CLI and Bash actions without supplying an install, binary, or concrete implementation — an incoherence that could cause the agent to attempt arbitrary shell operations or mislead users about how data is collected.

Purpose & Capability

The stated purpose (track ASINs, store snapshots, compute trends) aligns with the described files and outputs. However, the SKILL.md repeatedly references a 'spy' CLI (commands like `spy add`, `spy check`, `spy save`) while the skill package contains no binaries, no install spec, and no code — there is no explained mechanism that would provide the 'spy' command. That mismatch is not proportionate to the stated purpose and is an operational/integrity gap.

Instruction Scope

Instructions direct the agent to create and write files under ~/amazon-spy/ (watchlist.md, snapshots/, reports/, alerts.md) and to save snapshots. That storage is reasonable for a tracker. But because the skill allows Bash (allowed-tools: Bash) and references CLI commands with no implementation, an agent executing the skill could attempt arbitrary shell operations to implement the described behavior. The SKILL.md does not show explicit safe, limited commands or explain how data is obtained — it assumes user-provided listing data or unspecified automation, which is vague and grants broad discretion.

Install Mechanism

There is no install spec and no code files; this is instruction-only. From a supply-chain perspective that's lower risk because nothing external is automatically downloaded or written during install. However, the absence of an implementation for the referenced CLI is the main incoherence (see purpose_capability).

Credentials

The skill requests no environment variables, no credentials, and no config paths. That is proportionate to the written instructions which claim no API keys or scraping tools are required (the user supplies ASINs and/or listing data).

Persistence & Privilege

The skill writes persistent data to the user's home directory (~/amazon-spy/). always:false (not forced) and no special system privileges are requested, which is reasonable for a tracking tool. Note: autonomous invocation is allowed by default; combined with Bash access and the CLI-implementation gap, this could expand impact if the agent attempts to execute unvetted shell operations to 'implement' the CLI behavior.

What to consider before installing

This skill's high-level design (track ASINs, save snapshots under ~/amazon-spy/) is plausible and doesn't ask for secrets, but it references a 'spy' CLI and Bash usage without providing any code or install instructions. Before installing or using it: 1) Ask the publisher for the actual implementation (CLI binary, script, or install instructions) or a link to a release you can review. 2) Do not give sensitive credentials or paste private data into the tool until you can inspect what runs. 3) If you decide to run it, inspect the contents of ~/amazon-spy/ and any scripts the agent runs; consider running in a sandbox or container first. 4) If the skill claims to fetch remote data automatically, require that it specify the exact mechanism (API, authorized scraping tool, or manual paste). The assessment would become more confident and could move to 'benign' if the repository provides a signed release or clear installable CLI and the implementation only manipulates the declared workspace files without unexpected network/file access.

Track competitor ASINs over time and surface competitive intelligence — price shifts, BSR trends, review velocity, and listing changes.

Paste an ASIN or describe a competitor product. The agent tracks changes, computes trends, and alerts you when competitors make significant moves.

Commands

spy add <asin>                     # add an ASIN to your watchlist
spy check                          # run check on all tracked ASINs
spy compare                        # side-by-side competitor comparison matrix
spy price history                  # show price change timeline for tracked ASINs
spy bsr trend                      # 30-day BSR movement analysis
spy review velocity                # reviews/week rate for each tracked ASIN
spy listing diff                   # detect title/bullet/image changes vs. last snapshot
spy report                         # full competitive intelligence report
spy save                           # save current watchlist and snapshots to workspace

What Data to Provide

The agent works with:

• ASIN list — paste ASINs directly, one per line or comma-separated
• Verbal description — "my competitor is B08XYZ123, they just dropped price to $24.99"
• Listing data — paste competitor title, bullets, price, BSR rank from Amazon
• Historical notes — "last month their BSR was 3,200, now it's 980"

No API keys needed. No scraping tools required.

Workspace

Creates ~/amazon-spy/ containing:

• watchlist.md — tracked ASINs with metadata and notes
• snapshots/ — listing snapshots per ASIN (ASIN-YYYY-MM-DD.md)
• reports/ — generated competitive intelligence reports
• alerts.md — triggered alerts log

Analysis Framework

1. ASIN Tracking Setup

• Store ASIN, product title, brand, category, initial price, initial BSR, initial review count
• Record date added and last-checked timestamp
• Tag each ASIN with a competitive tier (direct / indirect / aspirational)
• Link your own ASIN to each competitor for direct comparison

2. Price Monitoring (Keepa-Style Logic)

• Track price changes over time with date-stamped entries
• Compute price volatility: (max price − min price) / avg price
• Flag: sustained price drops >10% — potential margin squeeze or launch push
• Flag: price increases >15% — possible supply issues or repositioning
• Identify coupon/promo patterns if mentioned (e.g., "20% off coupon visible")
• Compare competitor price vs. your price: gap analysis (you're $X cheaper/more expensive)

3. BSR Trend Analysis (30-Day Window)

• Log BSR data points at each check with timestamps
• Compute BSR velocity: (BSR start − BSR end) / days = ranks gained per day
• Positive velocity = ranking improving (lower BSR number)
• Flag: BSR improvement >500 ranks/week — aggressive launch or promotion detected
• Flag: BSR deterioration >1,000 ranks/week — possible listing suppression or inventory out
• Output sparkline-style trend table with direction indicators (↑↓→)

4. Review Velocity Analysis

• Track total review count at each snapshot date
• Compute weekly velocity: (reviews now − reviews at start) / weeks elapsed
• Industry benchmark: healthy launch pace = 10–30 reviews/week
• Flag: velocity >50 reviews/week — possible incentivized review campaign (watch for TOS risk)
• Flag: sudden review drop — potential review removals (negative signal)
• Compute projected total reviews in 30/60/90 days at current velocity

5. Listing Change Detection

• Snapshot: title, bullet points (all 5), A+ content presence, image count, video presence
• On spy listing diff, compare current snapshot vs. previous snapshot
• Highlight: title keyword changes (competitive repositioning signal)
• Highlight: bullet restructuring (conversion optimization attempt)
• Highlight: new images or video added (brand investment signal)
• Highlight: variation count changes (new sizes/colors = market expansion)

6. Competitive Positioning Matrix

• Build a comparison table: Your ASIN vs. each competitor
• Columns: Price | BSR | Reviews | Rating | Images | A+ | Video | Prime
• Score each competitor 1–5 on listing quality
• Identify your advantages and gaps
• Output recommended actions based on gaps found

Output Format

Every report outputs:

1. Competitive Snapshot — current standings table for all tracked ASINs
2. Movement Alerts — any significant changes since last check
3. Trend Charts — BSR and price trends (text-based sparklines)
4. Opportunity Flags — competitor weaknesses you can exploit
5. Recommended Actions — prioritized list of competitive responses

Rules

1. Always ask for the user's own ASIN when adding competitors — context requires knowing your position in the market
2. Never draw conclusions from a single data point — require at least 2 snapshots before declaring a trend
3. Flag data gaps explicitly — if a metric is missing, say so rather than estimating silently
4. Distinguish between correlation and causation — a BSR improvement could be organic or promotional
5. Never recommend matching a competitor's price cut without first checking your own margin floor
6. Save all snapshots to ~/amazon-spy/snapshots/ when spy save is called — tracking is only useful if historical data is preserved
7. Alert on anomalies proactively — if velocity data suggests a launch push or suppression event, surface it immediately