Competitor Monitor

When to use

Trigger this skill when the request includes any of:

• "Run the weekly competitor digest"
• "What did [competitor] ship this week?"
• "Did pricing change on any of our competitors?"
• "Summarize what's new across our competitive set"
• A scheduled Monday-morning job

Do not use for deep one-off competitor teardowns (that is a separate analyst engagement), win/loss analysis (sales-led), or feature-by-feature comparison tables (use content-writer with the competitor data as input).

Inputs

Required:

• Competitor list — 3-8 named competitors with URLs for: blog, newsroom/changelog, pricing page, and primary LinkedIn account.
• Our category — so the skill can judge what matters vs what is noise.

Optional:

• Previous digest — the last weekly output. Enables week-over-week diff framing.
• Focus areas — e.g., "we especially care about pricing changes and enterprise positioning."
• Cadence override — defaults to weekly (Monday 9am); biweekly or monthly can be requested.

Outputs

A single dated digest with these sections:

1. Header — week range (e.g., "Week of 2026-04-13 to 2026-04-19"), competitors tracked, and a one-line executive summary.
2. New content — per competitor, 1-3 bullet summaries of new blog posts, case studies, whitepapers, or videos shipped this week. Each with URL and publish date.
3. Positioning shifts — any change in homepage hero, tagline, category claim, or messaging hierarchy vs the previous snapshot.
4. Pricing changes — added/removed tiers, price deltas, new add-ons, changes to free tier or trial length.
5. Product moves — launches, major feature additions, integrations, deprecations, public beta announcements.
6. Social / executive signal — notable posts from competitor CEOs, heads of product, or official company accounts (LinkedIn and X).
7. Our recommended response — for each signal that matters, a specific action: update comparison page, draft rebuttal post, brief sales, ignore, or escalate.

Tool dependencies

• WebSearch / web-fetch (required) — for pulling blog posts, pricing pages, and newsroom updates.
• Optional: RSS reader or Changetower for automated change detection on pricing and landing pages.
• Optional: LinkedIn and X via whatever access is available (public pages only).
• Optional: Notion MCP to file the digest in the marketing intel workspace.
• Optional: Slack MCP to post the digest into #marketing on completion.

Procedure

1. Set the baseline

1. Load last week's digest if provided. If not, note that this is a baseline run and subsequent runs will be diff-oriented.
2. Confirm the competitor list and URLs. Flag any URL that has 404'd since last run — itself a signal.

2. Gather signals per competitor

For each competitor, check in this order (roughly 5-10 minutes per competitor if automated):

1. Blog / newsroom RSS. Pull everything published in the last 7 days.
2. Pricing page. Fetch current HTML; diff against last week's snapshot. Note any changed numbers, added tiers, or removed features.
3. Homepage hero. Fetch current copy above the fold. Note any change in H1, tagline, or primary CTA.
4. Product changelog or release notes. If public, pull entries from the last 7 days.
5. LinkedIn company page. Top 3 posts from the last 7 days, ranked by engagement.
6. Executive LinkedIn (CEO, CPO, CMO). Top posts from the last 7 days if public.
7. X/Twitter for the company handle and 1-2 key executives, if active.

3. Classify each signal

For every finding, tag it:

• Severity: high (material strategic move), medium (notable tactical move), low (routine activity).
• Category: content / positioning / pricing / product / social.
• Relevance to us: direct overlap / adjacent / unrelated.

Drop "low severity + unrelated" signals. They are noise.

4. Write the digest

• Keep each bullet to 1-2 sentences. This is a scannable document.
• Lead every section with the most important signal, not the chronologically first one.
• If a section has nothing material, write "No material activity." Do not pad.

5. Write the recommended response

For each high or medium signal, produce one of five responses:

• Update comparison page — when a competitor ships a feature we already match or beat.
• Draft rebuttal content — when positioning or messaging directly attacks our category claim.
• Brief sales — when pricing changes or a new tier affects active deals.
• Ignore — default for low-relevance moves. Explicit ignores save future cycles.
• Escalate — pricing undercut, direct feature parity with our moat, or an acquisition/funding event.

6. Close the loop

• File the digest (Notion, Google Doc, or wherever the team stores intel).
• Post a 3-bullet summary into #marketing on Slack with a link to the full doc.
• Save this week's page snapshots for next week's diff.

Examples

Example 1: Weekly digest

Output (abridged):

# Competitive digest — week of 2026-04-13 to 2026-04-19

Competitors tracked: CircleCI, GitHub Actions, Buildkite, Earthly, Depot.

**Executive summary:** Depot shipped a per-minute pricing change that undercuts
us by 15% on the smallest tier. Buildkite is quietly pivoting messaging from
"flexible CI" to "CI for platform teams" — direct overlap with our positioning.
Everything else is routine.

## New content

**Depot**
- "Why per-minute pricing is the only honest CI pricing" (blog, 2026-04-15). A
  direct shot at capacity-based pricing. URL: depot.dev/blog/per-minute
- Case study with Vercel's internal platform team (2026-04-17).

**Buildkite**
- "Buildkite for platform teams" (landing page, 2026-04-14). New segment page.
- "The paved road pattern in CI" (blog, 2026-04-18). Uses "paved road" language
  for the first time on their blog.

**GitHub Actions**
- No material content. Minor docs updates only.

**CircleCI**
- "2026 State of Software Delivery Report" (whitepaper, 2026-04-16). Annual
  report, mostly defensive.

**Earthly**
- No material activity.

## Positioning shifts

**Buildkite** — homepage H1 changed from "The hybrid CI/CD platform" to "CI
built for platform teams." Tagline and hero CTA also updated. This is a
category-level shift targeting the same buyer we target.

**Depot** — no homepage change, but new pricing-page subhead: "The only
per-minute CI that shows you the meter." Direct pricing-transparency play.

## Pricing changes

**Depot** — smallest tier dropped from $0.008/min to $0.005/min (-37%). Free
tier unchanged at 400 min/month. Enterprise pricing still "contact sales."

No changes on CircleCI, GitHub Actions, Buildkite, or Earthly.

## Product moves

**Depot** — shipped "Depot Policies" public beta (2026-04-17). Appears to be
a direct response to our Runners v2 sidecar. Feature parity on 3 of 5 capabilities
listed on our launch page; misses per-network-route policy and Kubernetes-native
enforcement.

**GitHub Actions** — `actions/cache@v4` GA. Minor.

**Buildkite** — no product moves.

## Social / executive signal

**Depot's CEO** — LinkedIn post on 2026-04-17 announcing Policies beta has
280+ reactions and 40+ comments. Several comments from our active prospects.

**Buildkite's Head of Product** — LinkedIn post framing "platform engineers
need tools, not platforms." Direct jab at the IDP/paved-road category.

## Our recommended response

| Signal | Response |
|---|---|
| Depot price cut -37% on smallest tier | **Escalate** to leadership + brief sales. Our smallest tier is now 20% more expensive. Decide this week whether to match, differentiate on value, or leave it. |
| Depot Policies beta | **Update comparison page** this week. Lead with the two capabilities they lack. Draft a short blog post (not a rebuttal, a technical deep-dive on why our approach handles the two cases they miss). |
| Buildkite repositioning to "CI for platform teams" | **Draft rebuttal content.** A 300-word LinkedIn post from our CEO clarifying the difference between "CI with platform features" and "a paved-road runtime." Not a takedown, a definition. |
| Buildkite's "paved road pattern" blog | **Ignore,** but note it. Our category language is becoming table stakes. |
| Depot Vercel case study | **Brief sales.** Vercel is a logo in our active pipeline. Sales needs this before the next call. |
| CircleCI annual report | **Ignore.** |

Example 2: Baseline (first-ever) run

Output (abridged):

# Competitive digest — baseline snapshot, 2026-04-19

This is a baseline run. Future digests will be framed as week-over-week diffs
against this snapshot.

Competitors tracked: CircleCI, GitHub Actions, Buildkite, Earthly, Depot.

**Executive summary:** Baseline only. Snapshots of pricing, homepage hero, and
latest 5 blog posts captured per competitor. See full doc.

## Current pricing snapshots

[Per-competitor table of tiers, prices, and free-tier terms as of 2026-04-19]

## Current homepage heroes

[Per-competitor H1 + tagline + primary CTA as of 2026-04-19]

## Latest blog posts

[Last 5 posts per competitor with URLs and dates]

## Our recommended response

No deltas yet. Next week's digest will surface changes against this baseline.
Suggested follow-up this week: subscribe to each competitor's RSS feed, set up
Changetower on each pricing page and homepage.

Constraints

• Factual accuracy over speed. Every claim in the digest is linked to a URL or explicitly flagged as "unverified."
• No speculation dressed as fact. If we think a competitor is planning a move, that is flagged as "hypothesis," not stated as news.
• No direct quoting of competitor customers unless the quote is already public.
• No accessing paywalled or gated content beyond what is publicly available. If a case study is gated, note it as gated and skip.
• Respect robots.txt and terms of service on all fetches.
• Digest length: 1-3 pages. Longer than that and nobody reads it. Cut ruthlessly.
• Every "recommended response" has an owner implied (sales, content, exec) and a timeframe (this week, this sprint, next quarter).
• Ignore is a valid response. Use it often. Not every competitor move deserves a reaction.

Quality checks

Before returning the digest, confirm:

• [ ] Every signal has a source URL or is explicitly flagged unverified.
• [ ] Executive summary is ≤3 sentences and front-loads the single most important finding.
• [ ] Each section has real content or says "No material activity" — no padding.
• [ ] Pricing changes are quantified (absolute and percentage).
• [ ] Positioning shifts include the old and new language side by side.
• [ ] Every high/medium signal has a specific recommended response.
• [ ] At least one "ignore" appears — if everything is urgent, nothing is.
• [ ] Digest fits in 1-3 pages when rendered.
• [ ] Competitor list matches the input (nothing dropped, nothing added silently).
• [ ] Baseline snapshots saved for next week's diff.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: tahirraufkeeyu
• Source: tahirraufkeeyu/software-development-agent-stack--sdas
• License: MIT

Install and usage instructions live in the source repository linked above.